Full text of "Studies in the Masoretic Tradition of the Hebrew Bible"

See other formats

Cambridge Semitic Languages and Cultures

Studies in the Masoretic
Tradition of the Hebrew Bible
Edited By

DANIEL J. CROWTHER, AARON D. HOoRNKOHL AND
GEOFFREY KHAN

a UNIVERSITY OF
9 CAMBRIDGE

Faculty of Asian and Middle
Eastern Studies

STUDIES IN
THE MASORETIC TRADITION
OF THE HEBREW BIBLE

Studies in
the Masoretic Tradition
of the Hebrew Bible

edited by
Daniel J. Crowther, Aaron D. Hornkohl
and Geoffrey Khan

UNIVERSITY OF
'9 CAMBRIDGE OpenBook
oe a Publishers LN)

ihe peering

https://www.openbookpublishers.com

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0
International (CC BY-NC 4.0). This license allows you to share, copy, distribute and
transmit the text; to adapt the text for non-commercial purposes of the text providing
attribution is made to the authors (but not in any way that suggests that they endorse you
or your use of the work). Attribution should include the following information:

Studies in the Masoretic Tradition of the Hebrew Bible, Daniel J. Crowther, Aaron D.
Hornkohl and Geoffrey Khan. Cambridge Semitic Languages and Cultures 15. Cambridge,
UK: Open Book Publishers, 2022, https://doi.org/10.11647/OBP.0330

Copyright and permissions for the reuse of many of the images included in this publication
differ from the above. Copyright and permissions information for images is provided
separately in the List of Illustrations.

Further details about Creative Commons licenses are available at https://creativecommons.
org/licenses

All external links were active at the time of publication unless otherwise stated and have
been archived via the Internet Archive Wayback Machine at https://archive.org/web

Updated digital material and resources associated with this volume are available at
https://doi.org/10.11647/OBP.0330#resources

Every effort has been made to identify and contact copyright holders and any omission or
error will be corrected if notification is made to the publisher.

ISBN Paperback: 9781800649194
ISBN Hardback: 9781800649200
ISBN Digital (PDF): 9781800649217
DOI: 10.11647/OBP.0330

Semitic Languages and Cultures 15.

ISSN (print): 2632-6906
ISSN (digital): 2632-6914

Cover images: A fragment of a Hebrew Bible manuscript (1 Sam. 25.44-26.8) from the
Cairo Genizah containing vocalisation, accents, Masoretic notes and Masoretic marks
(Cambridge University Library, T-S A8.10). Courtesy of the Syndics of Cambridge
University Library.

Cover design: Jeevanjot Kaur Nagpal

CONTENTS

CONTRIBUTORS vs. sieesvenucteestesevasvunedtaceveeaedsvcuat cevtewsocsseeedten ened vii
PREBA GE sacaisiict castrated siieedt on stie aves shen ston suicssdedSuesdtloasdewdtwadtlen seas dees xi
ABSTRACTS sous taceeevesedsnesetacele RivedaliercBeaeuesebuddiesdbuseteseadileseteaaeies Xv
Elvira Martin-Contreras

Using the Masora for Interpreting the Vocalisation

and Accentuation of the Biblical Text... ee eeee coos 1
Kim Phillips

The Masoretic Notes in RNL EVR II B 80+: An

MEAL REPO iy cic duces occau tative usta deaxteienteedieaiatnantiteacte nen 23

Vincent D. Beiler
The Marginal nun/zayin: Meaning, Purpose,
LOCALS AON ieiad.mnisarndeteiacraricaiininlad sandeedenptachsiincaexinyye 75

Aaron D. Hornkohl

Tiberian ketiv-gere and the Combined Samaritan
Written-Reading Tradition: Points of Contact and
COMUTASD ss ciisisuantisvsanchiaspasctisusanciaarssduasvesenssatsaeeuaseenanton 115

Estara J. Arrant

A Further Analysis of the ‘Byzantine (Italian-
Levantine) Triad’ of Features in Common Torah
GOGICES sestisstestecvonsatonstetiacuonast onsen Ute rensatioasentebeareesemuans 163

Geoffrey Khan

Hebrew Vocalisation Signs in Karaite Transcriptions
of the Hebrew Bible into Arabic Script...............:seeeee 203

vi Studies in the Masoretic Tradition of the Hebrew Bible

Yochanan Breuer

Dissonance between Masoretic Vocalisation and
Cantillation in Biblical Verse Division.....................0006+ 243

Daniel J. Crowther

Why are there Two Systems of Tiberian Te‘amim? ...... 289
Benjamin Williams

“Some Fanciful Midrash Explanation”: Derash on the

Te‘amim in the Middle Ages and Early Modern

POO se iarcisiceah eesdiariedsart anstdhareriiad eid ahead 329
Joseph Habib

Does Saadya Refer to the Accents in his Introduction
forthe Pemtate wel? si statics sda pinadstinecd deta dats cataaadas 377

CONTRIBUTORS

Elvira Martin-Contreras (PhD, Universidad Complutense, Ma-
drid, 2000) is Senior Research Fellow at the Institute of Lan-
guages and Cultures of the Mediterranean and the Near East (ILC-
CSIC). Her research is focused on the textual transmission and
the reception of the Hebrew Bible text attested in the rabbinic
literature and the Masora, i.e., the marginal annotations placed
in medieval Hebrew Bible manuscripts. She is also interested in
annotation practices in medieval Hebrew Bibles, as well as He-
brew palaeography. She is author of several monographs and
many articles, including Masora: La transmision de la Tradicion de
la Biblia Hebrea (Navarra, 2010). She is co-editor of The Text of
the Hebrew Bible: From the Rabbis to the Masoretes (Vandenhoeck
& Ruprecht, 2014).

Kim Phillips (PhD, University of Cambridge, 2016) is a Research
Associate in the Taylor-Schechter Genizah Research Unit, Re-
search Associate in Semitic Biblical Texts at Tyndale House, Cam-
bridge, and an affiliated lecturer in Hebrew in the Faculty of Di-
vinity of the University of Cambridge. His research focuses on
Masoretic Studies, medieval Jewish biblical exegesis, and Chris-

tian Palestinian Aramaic.

Vincent D. Beiler (PhD forthcoming, University of Cambridge,
2022) works on early Masoretic Bibles, with a special focus on
Bibles found in St Petersburg. His thesis is entitled ‘The Small

viii Studies in the Masoretic Tradition of the Hebrew Bible

Masorah: Genealogical Relationships in the Earliest Hebrew Bible

Codices’. This is his first publication.

Aaron D. Hornkohl (PhD, The Hebrew University of Jerusalem,
2012) is the author of Linguistic Periodization and the Language of
Jeremiah (Brill, 2013), a translated adaptation of his doctoral dis-
sertation, and co-editor (with Geoffrey Khan) of Studies in Semitic
Vocalisation (University of Cambridge and Open Book Publishers,
2020). He teaches Hebrew in the Faculty of Asian and Middle
Eastern Studies at the University of Cambridge. His recent re-
search focuses on the historical depth of the Tiberian reading tra-
dition as manifested in dissonance between its written and reci-
tation components, but he also deals with other ancient Hebrew
traditions, such as Qumran Hebrew and Samaritan Hebrew, dia-
chrony and linguistic periodisation, the Biblical Hebrew verbal

system, and pragmatics.

Estara J Arrant (PhD, University of Cambridge, 2021) is a Re-
search Associate at the Taylor-Schechter Genizah Research Unit
at the Cambridge University Library, working as the digital hu-
manities scholar for the ERC project TEXTEVOLVE: A New Ap-
proach to the Evolution of Texts Based on the Manuscripts of the Tar-
gums. Her research focuses on Semitic languages in contact, the
development and transmission of scriptures in Semitic languages,
Jewish and Islamic codicology and palaeography, and data sci-
ence and digital humanities. Her PhD thesis, ‘A Codicological and

Linguistic Typology of Common Torah Codices from the Cairo

Contributors ix

Genizah’, analysed the features of 1,500 everyday medieval To-

rah fragments using machine learning

Geoffrey Khan (PhD, School of Oriental and African Studies,
London, 1984) is Regius Professor of Hebrew at the University of
Cambridge. His research publications focus on three main fields:
Biblical Hebrew language (especially medieval traditions), Neo-
Aramaic dialectology, and medieval Arabic documents. He is the
general editor of The Encyclopedia of Hebrew Language and Lin-
guistics and is the senior editor of Journal of Semitic Studies. His
most recent book in the field of Hebrew is the two-volume The
Tiberian Pronunciation Tradition of Biblical Hebrew (University of
Cambridge and Open Book Publishers, 2020).

Yochanan Breuer (PhD, The Hebrew University of Jerusalem,
1993) is the Ch. N. Bialik Professor of the Hebrew Language at
the Hebrew University of Jerusalem and a member of the Acad-
emy of the Hebrew Language. His research interests include Bib-
lical Hebrew and the Masora, Mishnaic Hebrew, Babylonian Ara-
maic, and Modern Hebrew (especially the language of S. Y. Ag-
non). He is the co-editor of the series Language Studies (Jerusalem:
Mandel Institute). His most recent book is From Aramaic into He-
brew: The Method of Translation in the Book Hilkhot Re’u (The
Academy of the Hebrew Language, 2020).

Daniel J. Crowther (PhD, University of Bristol, 2015) is an As-
sistant Director for Langham Scholars’ Ministry and a Research

Associate of the Centre for Muslim and Christian Studies, Oxford.

x Studies in the Masoretic Tradition of the Hebrew Bible

He co-edited Reading the Bible in Islamic Context and has published
a number of articles in journals and collected volumes. His pre-
sent research concerns the Muslim historical setting of the Tibe-
rian Masoretic endeavour, the Ashkenazi handling of the Tiberian
heritage and the varied and various interconnections between the

Psalms and the Qur’an.

Benjamin Williams (PhD, University of Oxford, 2012) is Senior
Lecturer in Biblical and Rabbinic Studies at Leo Baeck College
and James Mew Lecturer in Rabbinic Hebrew at the University
of Oxford. His research focuses on the development of rabbinic
Bible interpretation, including the dating of late antique exegeti-
cal traditions; the transmission of midrash in manuscript and
print, particularly in the Ottoman Sephardi communities in
which the editiones principes were published; and the interpreta-
tion of midrash in Jewish and Christian commentarial traditions.
He is the author of Commentary on Midrash Rabba in the Sixteenth
Century: The Or ha-Sekhel of Abraham ben Asher (OUP, 2016).

Joseph Habib (PhD, University of Cambridge, 2021) is an inde-
pendent researcher. His doctoral thesis was entitled ‘Accents,
Pausal Forms and Qere/Ketiv in the Bible Translations and Com-
mentaries of Saadya Gaon and the Karaites of Jerusalem’. He cur-
rently works in the areas of medieval Jewish exegesis, Hebrew
accents, Judaeo-Arabic literature, and Hebrew philology. He has
published a number of articles on these topics in collected vol-

umes and journals.

PREFACE

This volume brings together papers on topics relating to the
transmission of the Hebrew Bible from Late Antiquity to the Early
Modern period. We refer to this broadly in the title of the volume
as the ‘Masoretic Tradition’. The term ‘Masoretic’ is sometimes
used in a narrower sense to refer to the activities of circles of
scholars known as Masoretes in the early Islamic period. The
most prestigious circle of Masoretes were those of Tiberias, who
produced some of the most authoritative medieval codices of the
Hebrew Bible, such as the Aleppo Codex (generally referred to by
the abbreviation A). The Tiberian Masoretes were associated with
the so-called Palestinian Yeshiva, which was the main seat of au-
thority in Palestine from Late Antiquity to the Middle Ages. We
have records of the activity of several generations of Tiberian
Masoretes from the eighth to the tenth centuries CE. By the sec-
ond half of the tenth century, the school of Masoretes in Tiberias
was discontinued for reasons that are not entirely clear.

The objective of the Tiberian Masoretes was the careful
preservation of the transmission of a stabilised form of the He-
brew Bible. They achieved this through the textualisation of the
oral reading that was received from antiquity in the form of
vowel and cantillation signs and the development of textual
notes, known as Masoretic notes. The Masoretic notes related to
differences in orthography, with statistical information about
their distribution, dissonances between orthography and oral
reading, and occasionally also differences in the interpretation of

words of similar form.

xii Studies in the Masoretic Tradition of the Hebrew Bible

The activities of stabilisation of the transmission of the He-
brew Bible, however, predate the formation of the Masoretic
school in Tiberias. Already in the Second Temple Period, author-
itative forms of both the written transmission and oral reading
tradition had begun to be fixed. Moreover, Masoretic activities
and the production of authoritative Masoretic Bible codices con-
tinued after the discontinuation of the Tiberian Masoretic School
in the tenth century. Indeed, the Codex Leningradensis (generally
referred to by the abbreviation L), which is the basis of modern
scholarly editions of the Hebrew Bible such as BHS and BHQ, was
produced in the eleventh century in Egypt. These Masoretic ac-
tivities continued in various centres in the Middle East and Eu-
rope down to the Early Modern Period.

Furthermore, despite the process of stabilisation, there has
always been some degree of diversity in the transmission of the
Hebrew Bible. This diversity can be seen in differences between
the authoritative oral reading tradition and the authoritative
written tradition, in differences in the systems of cantillation
across various parts of the Bible, and also in differences between
various written streams of transmission reflected by the extant
manuscripts. In the Middle Ages and beyond the written trans-
mission was more fixed, but minor differences, mainly in orthog-
raphy, are found across manuscripts. There were differences in
oral reading traditions and in systems of their textualisation.
There were also differences in the form and content of Masoretic
notes. Moreover, the engagement with the Masoretic tradition is

found in many rabbinic exegetical and grammatical works.

Preface xiii

The papers in this volume are studies on a range of aspects
of this Masoretic tradition of the Hebrew Bible in its broad sense,
ranging from the Second Temple Period to the Early Modern Pe-
riod. They focus on traditions of vocalisation signs and accent
signs, traditions of oral reading, traditions of Masoretic notes, as
well as rabbinic and exegetical texts.

We thank Estara Arrant and Vince Beiler, who helped
choose the images for the cover of the volume. We would like to
express our gratitude also to Open Book Publishers for all their
efficient help in publishing the volume. Their open-access initia-
tive will allow this publication to be widely read throughout the

world.

The Editors, Cambridge, September 2022

ABSTRACTS

Elvira Martin-Contreras, Using the Masora for
Interpreting the Vocalisation and Accentuation of
the Biblical Text

The marginal annotations that appear with the biblical text in
most medieval biblical manuscripts—called by the technical term
Masora—are hardly taken into account when interpreting the
biblical text. Their idiosyncratic characteristics (they are formu-
lated briefly, concisely, and, on many occasions, elliptically)
make it nearly impossible to appreciate the content of the anno-
tation and its possible interpretive relevance on a first reading.
All these difficulties can be resolved, however, by establishing
implicit information and formulating a clear methodology as to
how to analyse the Masoretic annotations. This allows us to study
them and apply them to the interpretation of the biblical text.
This article shows the benefits of using the Masora for the inter-
pretation of the biblical text through some selected examples, all
of them related to vocalisation and stress. The content of these
Masora annotations is explained and applied to textual interpre-

tation.

Kim Phillips, The Masoretic Notes in RNL EVR II B
80+: An Initial Report
RNL EVR II B 80 is a Torah codex, quintessentially Tiberian in

text and layout. Nonetheless, the masoretic notes reveal extensive

and sustained influence from the Babylonian masoretic tradition.

xvi Studies in the Masoretic Tradition of the Hebrew Bible

This influence can be detected in the technical terms employed
in the masoretic notes, the structure of the notes themselves, and
the biblical text-form implied by the content of the notes. This
article serves as a preliminary report demonstrating the nature
and extent of the Babylonian masoretic material in the manu-
script and illustrates some of the ways in which this material can
be used to consolidate and expand our existing knowledge of the

Babylonian Masora.

Vincent D. Beiler, The Marginal nun/zayin: Mean-

ing, Purpose, Localisation

In some early masoretic Bible codices, a large letter resembling
nun or zayin occurs in the margin, often in conjunction with the
marking of gere/ketiv. Occurring in some codices, but not in oth-
ers, the letter represents a bit of a cipher. Drawing on a database
of ca. 15,000 masora parva notes, taken from 81 different class-
marks, I propose that the letter, possibly a zayin, had (or ac-
quired) a practical purpose, viz. as a means of avoiding certain
types of copyist mistakes when recording gere/ketiv notes. Be-
cause the sign occurs in certain script types more than others, I
also show that the notation can function as something of a re-

gional identifier.

Aaron D. Hornkohl, Tiberian ketiv-qgere and the
Combined Samaritan Written-Reading Tradition:
Points of Contact and Contrast

Both the Tiberian and Samaritan biblical traditions are composite

in nature. In the Tiberian tradition this manifests most clearly in

Abstracts xvii

the phenomenon of ketiv-qere. Against the backdrop of the nor-
mally harmonious relationship between the written (i.e., conso-
nantal, orthographic) and pronunciation (i.e., vocalisation, reci-
tation) components of the Tiberian biblical tradition, ketiv-qere
instances are a clear indication of divergence between what is
written and what is read—divergence which, it should be empha-
sised, exceeds acknowledged cases of ketiv-qere. A similar rela-
tionship obtains between the Samaritan written tradition and its
oral recitation, with the latter regularly deviating from what was
evidently intended by the former. Both the Tiberian and Samari-
tan reading traditions are commonly characterised as later than
their respective written traditions. The present study examines a
series of ketiv-qere cases in the Pentateuch, seeking to explain the
various forms reflected by the Tiberian and Samaritan written

and reading traditions and to assess the relative antiquity of each.

Estara J Arrant, A Further Analysis of the ‘Byzan-
tine (Italian-Levantine) Triad’ of Features in

Common Torah Codices

This study analyses the distinctive features of a group of eleven
Torah fragments from the Taylor-Schechter collection of Cairo
Genizah manuscripts, which appear to come from related regions
and use the signs dagesh and shewa in three related ways to rein-
force a standard of pronunciation of the biblical text. The three
uses of these signs have, individually, been associated with Pal-
estino-Tiberian vocalisation, or labelled as ‘Extended Tiberian’. I
contribute a fresh analysis by contextualising the signs with each

other, showing how they work together to preserve a standard

xviii Studies in the Masoretic Tradition of the Hebrew Bible

form of pronunciation of the biblical text through reinforcing the
syllabification when the text is read aloud. I also examine the
codicological features of each of these fragments, which appear
very similar to each other. I conclude that they constitute a
group, and I infer what their physical and linguistic features re-
veal about their practical function in the reading and study of the

Hebrew Bible in the medieval period.

Geoffrey Khan, Hebrew Vocalisation Signs in
Karaite Transcriptions of the Hebrew Bible into

Arabic Script

In the 10th and 11th centuries CE many Karaite scribes in the
Middle East used Arabic script to write not only the Arabic lan-
guage, but also the Hebrew language. Such Hebrew texts in Ara-
bic transcription were predominantly Hebrew Bible texts. The
transcriptions reflect the oral reading tradition of the biblical
text. Most manuscripts reflect the Tiberian reading tradition.
Some reflect an imperfect performance of the Tiberian reading
tradition. This imperfect performance may be attributed to the
impact of the phonological system of the vernacular language of
the scribes. In this paper I discuss aspects of imperfect perfor-
mance discernible in the distribution of Hebrew vocalisation
signs that are used in the manuscripts. The paper focuses in par-
ticular on (a) deviations in the distribution of vowel signs that
reflect imperfect performance of Tiberian vowel qualities and (b)
deviations in the distribution of shewa and hatef signs that reflect

imperfect performance of Tiberian syllable structure.

Abstracts xix

Yochanan Breuer, Dissonance between Masoretic
Vocalisation and Cantillation in Biblical Verse
Division

The Masoretic text is the final stage of a process during which the
Masoretes had to decide between numerous various readings in
order to produce a fixed and consistent text. Although the final
production is a remarkable achievement, the Masoretic text still
contains cases of inconsistencies. The prominent example is the
discrepancy between the ketiv (the way the word should be writ-
ten) and the gere (the way the word should be pronounced),
where we find two contradictory readings in the same word. In
this article, a similar phenomenon is described regarding the vo-
calisation and the cantillation. Although the vocalisation and the
cantillation usually reflect division of a verse according to the
same interpretation, there are also cases where they reflect two
opposing divisions based on different interpretations. Awareness
of this may enrich our understanding of the complexity that was
involved in the fixing of the Masoretic text.

Daniel J. Crowther, Why Are There Two Systems

of Tiberian Te‘amim?

Why might it be that a dedicated system of accentuation is used
for ‘the Three’—the ‘poetic’ books of Job, Proverbs, and Psalms—
but not for the many other ‘poetic texts’ found scattered through-
out the ‘Twenty-One’ (the rest of the books of the Hebrew Bible)?

The earliest commentators associate the two types of Tiberian

Xxx Studies in the Masoretic Tradition of the Hebrew Bible

accentuation with differences in verse-length. More modern com-
mentators attribute it to the essence of poetry. Following these
two ideas, two different methods of presenting poetry can be ob-
served in the Twenty-One. One is appropriate to poetic texts with
short verses (of fewer than eight words per verse) and the other
is appropriate to poetic texts with long verses (of more than ten
words per verse). Within this double system, the practical chal-
lenges of presenting short-verse poetic texts under the accentua-
tion system of the Twenty-One can be observed in the one text
that attempts this feat (2 Sam. 22). This observation suggests a
rationale for a different system of accentuation that is more ap-
propriate to extended texts of exclusively short-verse poetry, as
found in the books of Psalms, Proverbs, and Job, but not in the

books of Chronicles, Lamentations and Song of Songs.

Benjamin Williams, “Some Fanciful Midrash Ex-
planation”: Derash on the Te‘amim in the Middle

Ages and Early Modern Period

This chapter examines the history of the idea that the shapes,
names, and sounds of the te‘amim convey information about bib-
lical narratives, including twists and turns in the plot, the
thoughts and motivations of the characters, and the way direct
speech was delivered. This exegetical technique is examined first
by enquiring into its relationship with the midrashic method of
deriving such information from the graphic features of the con-
sonantal text of the Hebrew Bible. Turning to the approach of

Tobias ben Eliezer, Joseph ibn Caspi, and Bahya ben Asher, at-

Abstracts xxi

tention is focused on interpretations of unusual and irregular can-
tillation marks, including the shalshelet, according to the princi-
ples of derash. Finally, examples from the commentaries of Moses
Alsheikh of Safed are examined to show how sixteenth-century
Sephardi interpreters treated the Masoretic system of accentua-
tion more broadly as a source of information concerning biblical

narratives.

Joseph Habib, Does Saadya Refer to the Accents in

His Introduction to the Pentateuch?

In the introduction to his long commentary on the Pentateuch,
the Rabbanite scholar Saadya Gaon discusses the importance of
word groupings. The possibility has been raised that here Saadya
is referring to the biblical accents. The purpose of this article is
to determine whether or not Saadya has the accents in mind. This
is done through a close analysis of select key terms and the bib-

lical passages mentioned in the passage.

USING THE MASORA FOR
INTERPRETING THE VOCALISATION
AND ACCENTUATION OF THE BIBLICAL
TEXT"

Elvira Martin-Contreras

Strictly speaking, the term Masoretic Text (MT)? refers to any He-
brew biblical codex that is accompanied by a corpus of marginal
annotations known as masora.? Each codex has its own set of mar-

ginal annotations and there are no two masoras that are the same

' This article was completed under the auspices of a research project
entitled ‘Legado de Sefarad II. La producci6n material e intelectual del
judaismo sefardi bajomedieval’, which is based at the ILC-CSIC in Ma-
drid and funded by the Plan Nacional de I+ D +i (FFI2015-63700-P).

? For the use of the term, see Martin-Contreras (2016, esp. 420).

In this paper the terms masora and Masorah are used according to the
distinction made by Aron Dotan. He divided written Masorah into two
categories: (1) the masoretic notes in the margins of the text and the
longer lists which accompany the text or are appended to it—the masora
in the narrow sense; (2) the graphemes which, by their nature, are of
two types: (a) vocalisation signs; (b) accentuation signs. See Dotan
(2007, 614). The term is written Masorah (with uppercase M and final
h) when it is the generic name, and masora (with lower case m) when
it refers to the marginal masoretic annotations of a particular manu-
script.

2 Martin-Contreras

(Orlinsky 1966, esp. xxxvi). The marginal annotations are found
in the intercolumn, top and bottom margins of each folio, and
also collected at the end of the biblical books, where they are
arranged in lists.* All of them contain varied information about
the words of the biblical text with which they appear, such as:
spelling, enumeration, vocalisation, accentuation, grammatical
rules, meaning, etc. (see Martin-Contreras 2021, 178-81). How-
ever, all this information is rarely taken into account when the
biblical text is interpreted.°

The roots of ‘neglecting’ the interpretative value of these
marginal annotations lie in (a) lack of knowledge about this
source (additional specialised training is needed to decipher the
annotations), (b) the way the annotations work, and (c) how the

information is provided.

These textual annotations are found in manuscripts vocalised in the
three systems of Hebrew vocalisation. Consequently, it is possible to
distinguish three kinds of Masoras: Tiberian, Palestinian, and Babylo-
nian. On Palestinian Masora see Kahle (1959); Weil (1963, 68-80);
Yeivin (1963); Revell (1970; 1974; 1977); Chiesa (1978). For a general
view of Babylonian Masora see Ofer (2001). An additional masora is
attached to the text of Targum Onkelos; see Klein (2000).

4 This information is sometimes denominated Masora Finalis; however,
as the Masoretic material arranged by Jacob ben Hayyim at the end
of the Second Rabbinic Bible is called Masorah Finalis, it is better to
avoid this term.

° For the benefits of using the masora for interpretation, see Freedman
and Cohen (1974); Fernandez Tejero (1984); Barthélemy (1992, lxix—
xcvii); Mynatt and Crawford (2001); Martin-Contreras (2009; 2013);
Dotan (2010).

Using the Masora for Interpreting Vocalisation and Accentuation 3

Usually, each marginal annotation is linked to one or more
words of the biblical text written on the same folio. A graphic
symbol—a small circle called a circellus (")—is often placed over
a word or between two or more words of the biblical text. The
circellus alerts us to the presence of extra information on the word
to be found in a marginal annotation.°

Annotations are connected to their lemmas through (a)
their placement next to the line of the text (this is the case with
annotations placed in the intercolumnar margins, which are
called collectively masora parva, MP) and (b) the repetition of the
lemma in the annotation itself. This latter technique is typically
used for those annotations written in the top and bottom margins
(all of which are called collectively masora magna, MM), as well
as for annotations found at the end of a biblical book or a collec-
tion of biblical books.

The denominations masora magna and masora parva merely
express an external-technical division of the annotations. This di-
vision does not imply differences in the function and nature of
the annotations placed in each. Both types annotate the same
kinds of information, but they differ in how they represent this
information in writing. It has been said that the MM can be re-
garded as an expansion of the information that is collected in the
MP. This is only partially true. There are many MP annotations

with no parallel MM, and vice versa. Therefore, it is better to

® Alternatively, a lemma may have a circellus, but no corresponding an-
notation; or, conversely, a lemma may have an annotation, but no cor-
responding circellus.

4 Martin-Contreras

regard both types, MP and MM, as parallel entities (see Dotan
2010, 59, n. 9).

The masoretic annotations are characterised by (a) their ex-
pression of information in a mixture of Rabbinic Hebrew and Ar-
amaic (Hyvernat 1902-1905) and (b) their brief and concise
presentation of information (generally using abbreviations) that
is, on many occasions, even elliptical (part of the information re-
mains implicit). There is no standardised form for these abbrevi-
ations, or a single way of expressing similar information: they
vary between manuscripts and, sometimes, even within the same
manuscript (see Fernandez Tejero and Ortega Monasterio 1981;
1983; Martin-Contreras 2012; Ortega Monasterio 1986; 1993;
1997; Fernandez Tejero 2009). Those placed in the intercolumn
margins show the briefest form, with the words often represented
only by their initial letters. The ultimate expression of this ellipsis
are annotations that give only a number (a letter of the Hebrew
alphabet with a supralinear dot).

In most cases, these characteristics make it impossible on
first reading to appreciate the content of the annotation and its
possible relevance to interpretation. However, all these difficul-
ties can be resolved by supplying the information that was left
implicit and by formulating a clear methodology of how to ana-
lyse masoretic annotations. This enables us to apply the infor-
mation they contain to the interpretation of the biblical text
(Martin-Contreras 2013).

Once the apparent difficulties posed by the Masora have

been explained, the best way to learn about its benefits for the

Using the Masora for Interpreting Vocalisation and Accentuation 5

interpretation of the biblical text is to use it. The following se-
lected examples show how to decipher the content of a masoretic
annotation so as to apply it thereafter to textual interpretation.

All of these examples concern vocalisation and accentuation.

1.0. Judg. 6.37

The following information on the word ann ‘dryness’ (Judg. 6.37)
is given in the masora of Leningrad codex (L).

Figure 1: Leningrad codex, f. 140r (courtesy of The National Library of
Russia)

siberian

marie pra

ENSTWADA AY

i ‘- .~" : t

Geaklrs Sigs — & . &. ss Ney wert = ae

The MP annotation says: ? ‘sixteen’. There is no annotation in the
MM. At first glance, the annotation could be classified as one of
numerical type, stating the number of times the word appears in
the Bible. However, a concordance search reveals that this word
appears eleven times (Even-Shoshan 1996, 398). Is the annota-
tion wrong? Or does it give information of a different type?
According the methodology to be followed in the analysis
of any masoretic annotation, the next step is to confirm the reli-
ability of the information given in the annotation. For this pur-
pose, it is necessary to consult the masoras in the main Tiberian
biblical manuscripts (B.L. Or. 4445, the Cairo codex of the Proph-
ets, the Aleppo Codex, and the ‘Leningrad Codex’) and the major

6 Martin-Contreras

masoretic lists and treatises (Frensdorff 1864; Dotan 1967; Diaz
Esteban 1975; Ginsburg 1975; Ognibeni 1995). I have searched
in the Cairo (C) and Aleppo (A) codices, but none of them have
masoretic annotations on this word in this verse. I then searched
in Ginsburg’s masoretic compilation. There is one list where the
lemma and the information match the annotation here (Ginsburg
1975, 497-98).

We have all the information collected in the process of an-
alysing the annotation: the MP information, the MM information,
the identification of the masoretic signs (simanim), other evi-
dence. This can help to explain the overall meaning of the anno-
tation and its purpose.

According to the list in Ginsburg’s compilation, the sixteen
references are: Gen. 31.40; Judg. 6.37, 39, 40; Jer. 36.30; 49.13;
50.38; Ezek. 29.10; Isa. 4.6; 25.4; 25.5, 5; 61.4; Zeph. 2.14; Hag.
1.11; and Job 30.30. After careful examination of the references,
we can infer two facts. First, the number sixteen includes occur-
rences of this word both with and without prefixes. This addi-
tional information is stated explicitly in the MP annotation on
this word at Job 30.30 in L: wa * ‘sixteen in the meaning’. Sec-
ond, the word is vocalised with segol under the resh in all the
instances. In other words, there are sixteen occurrences of the
word 15m and similar forms vocalised with segol.

But, why is it necessary to provide this information? Be-
cause the word 17n, with and without prefixes, also appears vo-

calised with sere in the Bible: 29h ‘Horeb’.’ The purpose of the

” Sixteen times in the Hebrew Bible plus one case where the word is
written plene, 19in (Exod. 33.6); cf. Even-Shoshan (1996, 352).

Using the Masora for Interpreting Vocalisation and Accentuation 7

annotation is to distinguish these consonantal homographs with
different meanings. The distinction is made via the vocalisation
and accentuation: those with segol and penultimate stress are
cases of the common noun 17h ‘drought, parching heat, desola-
tion/dryness’ (Brown 1952), and those with sere and stress on the

ultima are instances of the proper name 29h ‘Horeb’.

2.0. Zech. 6.10

The following information on the word nx; from Zech. 6.10 is

found in the masora of C.

Figure 2: Cairo codex, Zech. 6.10 (photographs held by the Masora team
at the CSIC)

The MP annotation says: 01 t. There is also a MM annotation:

nwdwi moana nNqpi ADA Onan wnwi nm pm pain’. por 7 mx
25137 nxn mip> JRATIAN

The textual information comes after the lemma and it explicitly
says we-simanehon. This is the introductory formula for saying
that the next words are the simanim, the catchwords that make it
possible to identify the verses involved. I have identified them as:
Gen. 6.18; Exod. 3.18; Deut. 17.9; 26.3; Jer. 36.6; 1 Sam. 20.19;
and Zech. 6.10.

8 Martin-Contreras

Similar information is found in the MP annotation on this
word at Zech. 6.10 in L and in Ginsburg’s Masoretic compilation
(Ginsburg 1975, 167). All this information is going to help us to
understand the annotation, the next step.

So, what does this annotation mean? Firstly, we translate
the MP annotation v1 ¢ ‘seven times with the accent’. According
to this translation, the first hypothesis is that the annotation re-
lates to the accent in the word, pashta; in other words, to the
seven times that the word appears with this accent. But if we
check the word in each of the verses given in the MM annotation,
we can see that this accent does not occur in all of them. It is
therefore necessary to pursue other clues.

The list in Ginsburg’s compilation adds a very important
piece of information: the word is accented these seven occur-
rences on the letter taw.

According to the concordances (Even-Shoshan 1996, 154),
the word nx, the 2ms gal perfect with the prefix waw, appears
nineteen times in the entire Bible. A careful examination of the
references confirms that: the word nxai in the seven verses listed
in the masoretic annotation has the accent on the ultima, and in
the other twelve instances on the penultima. In other words, the
meaning of the annotation is that the word occurs seven times
with stress on the ultima.

But, why is it necessary to give this information? What is
the purpose of the annotation? Is it merely statistical? Those who
think that the masora has a numerical character may answer ‘yes’.
But, my answer is ‘no’. The position of the word stress is often

used to distinguish similar words with different tense meanings:

Using the Masora for Interpreting Vocalisation and Accentuation 9

in the case of nxai, those with an ultima stress usually have future
meaning—the so-called waw consecutive perfect—while those
with penultima have past meaning (Revell 1985; Khan 2000, 92).
So, the purpose of the annotation is to ensure that the word is

not interpreted as a past tense form in the seven relevant verses.

3.0. Josh. 2.3

The word *xxin ‘bring out’ (Josh. 2.3) has the following MP an-

notation in A.

Figure 3: Aleppo codex, Josh. 2.3 (courtesy of the Ben-Zvi Institute, Je-
rusalem. Photographer: Ardon Bar Hama)

Site ssn pivaen oP ee

inet Ps 5
ot PUNE wag apy 6 hee Dee
There is no MM annotation. Similar information is found in the
MP annotation on this word in L.
What does this annotation mean? It can be translated as:
‘unique® in feminine; penultimate stress’. However, the word has

the accent telisha gedola and the sign of this accent is not usually

8 | prefer to translate the term n° let as ‘unique’ because it may refer
to words or expressions that appear once in the Bible (hapax in sensu
strictu) as well as to words or expressions that are unique in some other
sense (spelling, vocalisation, accentuation, meaning, location, etc.).

10 Martin-Contreras

used to indicate stress position (Yeivin 1980, 102). Codex L helps
to elucidate this matter.

The word in that codex has two signs of this accent, one
over the letter he and other over the letter sade: »»°¥i7.

Figure 4: Leningrad codex, f. 122r (courtesy of The National Library of
Russia)

Pom
rat angry

The sign over the letter sade is not reproduced in the standard

a1 bie

printed edition or the electronic ones (such as Bible Works and
Accordance), with the exceptions of the Biblia Hebraica Lenin-
gradensia (Dotan [ed.] 2001, 318) and the module ‘Masora Tesau-
rus’ (Dotan [ed.] 2014), both edited by Aron Dotan. This is one
of the reasons it is advisable always to check the manuscript
when working on the masora. The editions do not always offer all
the details exactly as they appear in the manuscripts.

The telisha sign is generally not repeated on the stressed
syllable in manuscripts with standard Tiberian pointing, alt-
hough there are some exceptions, and this case is one of them
(Yeivin 1980, 211). And, what does the repetition mean? Accord-
ing to Israel Yeivin (1980, 102), the sign is repeated on this word

to indicate the stress position.

Using the Masora for Interpreting Vocalisation and Accentuation 11

And, what is the purpose of the annotation? The word *x°xin
appears three times in the Bible: Jer. 7.22 (qere); 11.4; and here.’
The word is stressed on the ultima in the two verses in Jeremiah.
This information is confirmed by a list from the masoretic com-
pendium ’Okla we-’Okla, according to the Paris MS version, on
words that occur once with penultimate stress while everywhere
else they have ultimate stress (Frensdorff 1864, 171, 372). The
ultimate purpose of the annotation is to distinguish homographs.
The position of the word stress is used to do this: those with ulti-
mate stress are infinitives with the 1cs suffix, while the one with

penultimate stress is a FS imperative.

4.0. Deut. 32.5

An analysis of the accents and masora of Deut. 32.5 in L illustrates
its role and importance in interpreting the biblical text. This is a
difficult text, with a great variety of renderings attested in the
ancient versions. The accentuation of the first part of the verse
is: ow) ra x» 1) nnwW. According to the accents, x9 15 should be
read together and a literal translation of the first three words
would thus be—as odd as it may sound—‘they behaved corruptly
towards him not’. This accentuation is unusual and the trend has,
therefore, been to disregard it and follow the more ‘logical’ read-
ing proposed by the consonantal text. Accordingly, the transla-
tion found in the New American Standard Bible is: ‘They have

acted corruptly toward Him, They are not His children, because

° There is one further case where the word is written defectively, i.e.,
without the yod in the second syllable, Jer. 34.13; cf. Even-Shoshan
(1996, 484).

12 Martin-Contreras

of their defect’. However, the reading proposed by the accents is
supported by the Targumim and the masora. What does the ma-
sora teach us about this?

There is a MP annotation to x5 #5 in L:

Figure 5: Leningrad codex, f. 118v (courtesy of The National Library of
Russia)

Vo VY Gof ids. ban va Ly Va Joe
Sa | MRRU RNR Gale
iN non 4 . . OAT t.: 4

aS sah ete wha te

It says: 1 j ‘six [times] in the Torah’. There is no MM note. Man-
uscripts Or 4445 (B) and A have no masoretic annotation here,
but according to a list in Ginsburg’s (1975, 120) compilation the
references could be: Gen. 28.1; 47.18; Exod. 28.32; Deut. 23.17;
25.5; 32.5. Apart from the case in Deut. 32.5, these two words
do not appear to be linked conjunctively by the accents in any of
the remaining occurrences. It seems that this annotation merely
states the number of times that these two words appear together,
in this order, irrespective of the accentuation.'°

However, I have found others annotations in L which may
be relevant to the accentuation. The sequence ‘5 X4 in Gen. 38.9

has two marginal annotations:

10 This sequence is found also in Deut. 21.16.

Using the Masora for Interpreting Vocalisation and Accentuation 13

Figure 6: Leningrad codex, f. 23v (courtesy of The National Library of
Russia)

seen AAP TIMER He Ga agar arn oH Snot ch 300 ab gre
+ SPOT VINE Heng pie [a

barley tntroe ba on

Ay at vo

Piaiiniiani avo i > : chain! i
The MP annotation says: 7 ‘five’. Additional information and the
references are given in the MM annotation:

x> miawn nwd 1b xd naan cin taro 8902 pn pT AD ND

% nnw adn tm aan 1 xd) tayn xd1 > xd a Sy raynn ray b
n>

1 x5 five [times]: 1 x5 °D 7nX yT™ (Gen. 38.9); 7 TNX
1% x5 mann (Hab. 1.6); 1b xb miawn nw (Hab. 2.6); rap
95 x5 2 Sy aaynn (Prov. 26.17); Wan 1 xd1 Jnayn xd (Dan.
11.17); and one [with the order] reversed, 8515 nnw (Deut.
35.2).

Careful examination of the references allows us to infer that in

all of them these two words appear side by side, in either order,

14 Martin-Contreras

and linked—twice by the accents merkha and tarha and four times
by maqgef. So, these two words should be read together."

This MM annotation not only confirms the unusual accen-
tuation of the sequence in Deut. 32.5, but also demonstrates the
Masoretes’ concern that it be interpreted correctly. By listing this
case together with the occurrences that have maqgef,'* they en-

sured that it would be treated in the same way.

5.0. Judg. 5.8

The word on) ‘fighting, war’ (Judg. 5.8) has a MP annotation in
A.

EAR aES pe

PATTIE

Jerusalem. Photographer: Ardon Bar Hama)
i “ye \ .
thet
+ "
al

Figure 7: Aleppo codex, Judg. 5.8 (courtesy of the Ben-Zvi Institute,
iy aint Pw tH
pied ae NOVTRD nyt Roce me >
The same information is given in the MP annotations of codices
C and L to the same verse.
According to this annotation, the word is ‘unique’. How-
ever, we find in the concordances that this word appears 38 times

in the entire Bible.!? What, then, does the annotation mean? In

" For a different understanding of the Masoretic notes and of the verse
see McCarthy (2002).

” On the use of magqgef with words consisting of a single open syllable
see Yeivin (1980, 230-31).

3 Omitting occurrences of this word with prefixes; cf. Even-Shoshan
(1996, 596-97).

Using the Masora for Interpreting Vocalisation and Accentuation 15

what sense is the word is unique? List 373 of the masoretic com-
pendium ’Okhla we-’Oklah, according to the Paris ms version,
helps us with this dilemma (cf. Frensdorff 1864, 172). This is one
of the words that occur once with an ultimate stress while every-
where else they have a penultima stress.

By drawing our attention to the unusual stress position, the
annotation supplies the clue to properly understand this word in
context.’ It tells us that this is not another instance of the word
on) ‘bread’ in the pausal form on?—which is vocalised identically
to our word in Deut. 32.5, but with penultimate stress—but a
different word. The distinction between the two words is made
by the stress position. So the essential purpose of this annotation
is to avoid misunderstanding of the word as ‘bread’. Here the

word means ‘war’, from the root o"nd ‘fighting’.

6.0. Conclusion

To conclude, I hope these examples help to show why the masora
is an indispensable tool in our attempts to achieve a more pro-
found understanding of the Hebrew biblical text. It constitutes a
historical record that accompanies the biblical text: a window

onto the past. Why not use it?

* On the difficulty of interpreting this word in context, cf. Fernandez
Marcos (2011, 56*-57*).

16 Martin-Contreras

References

Barthélemy, Dominique. 1992. Critique textuelle de l’AT, vol. 3.
Orbis Biblicus et Orientalis 50/3. Fribourg: Editions Uni-
versitaires and Gottingen: Vandenhoeck & Ruprecht.

Brown, Francis, Samuel Rolles Driver, and Charles A. Briggs.
1906. A Hebrew and English lexicon of the Old Testament:
With an appendix containing the Biblical Aramaic. Based on
the Lexicon of William Gesenius as translated by Edward Rob-
inson. Oxford: Clarendon Press.

Chiesa, Bruno. 1978. L’Antico Testamento Ebraico secondo la
tradizione palestinense. Turin: Bottega d’Erasmo.

Diaz Esteban, Fernando. 1975. Sefer Oklah we Oklah. Coleccién de
listas de palabras destinadas a conservar la integridad del texto
hebreo de la Biblia entre los judios de la Edad Media. Textos y
estudios “Cardenal Cisneros” 4. Madrid: CSIC.

Dotan, Aron. 1967. Sefer Diqduge ha-Te‘amim of Aharon ben Moshe
ben Asher, with a Critical Edition of the Original Text from New
Manuscripts. Jerusalem: The Academy of the Hebrew Lan-

guage.

(ed.). 2001. Biblia Hebraica Leningradensia: Prepared Ac-
cording to the Vocalization, Accents, and Masora of Aaron ben
Moses ben Asher in the Leningrad Codex. Peabody, MA: Hen-
drickson.

. 2007 (1971). ‘Masorah’. In Encyclopaedia Judaica, 2nd
edition, 23: 603-56.

Using the Masora for Interpreting Vocalisation and Accentuation 17

. 2010. ‘Masora’s Contribution to Biblical Studies: Revival
of an Ancient Tool’. In Congress Volume Ljubljana 2007, ed-
ited by André Lemaire, 57-69. Vetus Testamentum Supple-
ments 133. Leiden: Brill.

(ed.). 2014. Masora Thesaurus. Accordance electronic edi-

tion. Altamonte Springs, FL: OakTree Software, Inc.

Even-Shoshan, Avraham. 1996. A New Concordance of the Bible.
Thesaurus of the Language of the Bible Hebrew and Aramaic
Roots, Words, Proper Names, Phrases and Synonyms. Jerusa-
lem: Kiryat Sefer.

Fernandez Marcos, Natalio (ed.). 2011. The Biblia Hebraica
Quinta, vol. 7, Judges. Stuttgart: Deutsche Bibelgesellschaft.

Fernandez Tejero, Emilia. 1984. ‘Masora y Exégesis’. In Simposio
Biblico Espafiol, edited by Natalio Fernandez Marcos, Julio
C. Trebolle Barrera, and Javier Fernandez Vallina, 183-91.

Madrid: Universidad Complutense.

. 2009. ‘Se equivocé el masoreta. ;Se equivocaba?’. Se-
farad 69: 303-13;

Fernandez Tejero, Emilia, and Maria Teresa Ortega Monasterio.
1981. ‘Las masoras de A, C y Len el libro de Nahum’. Sefa-
rad 41: 1-43.

. 1983. ‘Las masoras de A, C y L en el libro de Joel’. In

Estudios Masoréticos (V Congreso de la IOMS). Dedicados a

Harry M. Orlinsky, edited by Emilia Fernandez Tejero, 205-

42. Textos y estudios “Cardenal Cisneros” 33. Madrid: In-

stituto “Arias Montano”.

18 Martin-Contreras

Freedman, David Noel, and Miles B. Cohen. 1974. ‘The Masoretes
as Exegetes: Selected Examples’. In 1972 and 1973 Proceed-
ings IOMS, edited by Harry Orlinsky, 35-46. Missoula, MT:
Scholars Press.

Frensdorff, Solomon. 1864. Ochlah W’ochlah. Hannover: Hahn.

Ginsburg, Christian David. 1966 (1897). Introduction to the Mas-
soretico-Critical Edition of the Hebrew Bible. New York: Ktav
Publishing House; originally published London: Trinitarian
Bible Society.

. 1975. The Massorah Compiled from Manuscripts Alphabeti-

cally and Lexically Arranged. With an Analytical Table of

Contents and Lists of Identified Sources and Parallels by A.
Dotan. 4 vols. New York: Ktav Pub. House.

Hyvernat, Henri. 1902-1905. ‘Petite Introduction a l’Etude de la
Massore’. Revue Biblique 11: 551-63; 12: 529-42; 13: 521-
46; 14: 203-34, 515-42.

Kahle, Paul E. 1959. The Cairo Geniza. 2nd edition. Oxford: Black-
well.

Khan, Geoffrey. 2000. The Early Karaite Tradition of Hebrew Gram-
matical Thought. Studies in Semitic Languages and Linguis-
tics 32. Leiden: Brill.

Klein, Michael L. 2000. The Masorah to Targum Ongelos: As Pre-
served in Mss Vatican Ebreo 448, Rome Angelica Or. 7, Frag-
ments on the Cairo Genizah in Earlier Editions by A. Berliner
and S. Landauer—Critical Edition with Comments and Intro-

duction. Binghamton, NY: Targum Studies.

Using the Masora for Interpreting Vocalisation and Accentuation 19

Martin-Contreras, Elvira. 2009. ‘Masoretic and Rabbinic Lights
on the word ?17, Ruth 3:15: 17 or x12?’ Vetus Testamentum
59: 244-56.

. 2012. ‘The Phenomenon Qere we la’ ketib in the Main Bib-

lical Codices: New Data’. Vetus Testamentum 6: 77-87.

. 2013. ‘Masora and Masoretic Interpretation’. In The Ox-
ford Encyclopedia of Biblical Interpretation, edited by Steven
L. McKenzie, I: 542-50. Oxford: Oxford University Press.

. 2016. ‘Medieval Masoretic Text: Overview Article’. In The
Textual History of the Bible, edited by Armin Lange, 1A:
420-29. Leiden: Brill.

. 2021. ‘Annotations in the Earliest Medieval Hebrew Bible

Manuscripts’. In Scribal Habits in Near Eastern Manuscript
Traditions, edited by George A. Kiraz and Sabine Schmidtke,
167-88. Piscataway, NJ: Gorgias Press.

McCarthy, Carmel. 2002. ‘Masoretic Undertones in the Song of
Moses’. Proceedings of the Irish Biblical Association 25: 42-
45.

Mynatt, Daniel S., and Timothy G. Crawford. 2001. ‘Integrating
the Masorah into the Classroom: A Tribute to Page Kelley’.
Perspectives in Religious Studies 28: 373-79.

Ofer, Yosef. 2001. The Babylonian Masora of the Pentateuch its
Principle and Methods. Jerusalem: Magnes Press.

Ognibeni, Bruno. 1995. La seconda parte del Sefer Oklah weOklah.
Edizione del ms. Halle, Universitdtsbibliothek Y v 4°, ff. 68-
124. Textos y estudios “Cardenal Cisneros” 57. Madrid:
CSIC.

20 Martin-Contreras

Orlinsky, Harry M. 1966. ‘Prolegomenon. The Masoretic Text: A
Critical Evaluation’. In Ginsburg 1966 (1897), i-xlv.
Ortega Monasterio, Maria Teresa. 1986. ‘Las masoras de A, C y L

en el libro de Habacuc’. Henoch 8: 149-84.
. 1993. ‘Some Aspects of the Masora of the Codices Or.
4445 and Aleppo’. In Estudios Masoreticos (X Congreso de la

International Organization for Masoretic Studies), en memoria

de Harry M. Orlinsky, edited by Emilia Fernandez Tejero

and Maria Teresa Ortega Monasterio, 89-98. Madrid: Insti-

tuto de Filologia del CSIC.

. 1997. ‘Some Masoretic Notes of Mss. L and Or 4445 Com-
pared with the Spanish Tradition’. Sefarad 57: 127-33.

Revell, E. J. 1970. Hebrew Texts with Palestinian Vocalization. To-

ronto: University of Toronto Press.

. 1974. ‘The Relation of the Palestinian to the Tiberian
Massora’. In 1972 and 1973 Proceedings IOMS, edited by
Harry Orlinsky, 87-98. Masoretic Studies 1. Missoula: MT:

Scholars Press.
. 1977. Biblical Texts with Palestinian Pointing and their Ac-

cents. Masoretic Studies 4. Missoula, MT: Scholars Press.

. 1985. ‘The Conditioning of Stress Position in Waw Con-
secutive Perfect Forms in Biblical Hebrew’. Hebrew Annual
Review 9: 277-300.

Weil, Gérard E. 1963. ‘Un fragment de Okhlah Palestinienne’. An-
nual of the Leeds University Oriental Society 3: 68-80.
Yeivin, Israel. 1963. ‘A Palestinian Fragment of Haftaroth and

Other MSS with Mixed Pointing’. Textus 3: 121-27.

Using the Masora for Interpreting Vocalisation and Accentuation 21

. 1980. Introduction to the Tiberian Masorah. Translated and
edited by E. J. Revell. Masoretic Studies 5. Missoula, MT:

Scholars Press.

THE MASORETIC NOTES IN RNL EVR II B
80+: AN INITIAL REPORT

Kim Phillips

From Second Temple times until the end of the first millennium
CE, the Hebrew Bible was preserved and transmitted in two major
Jewish cultural centres: Tiberias and ‘Babylon’ (i.e., Iraq) (Khan
2020, 6-33). The consonantal texts preserved in each of these
two centres differed in a great many small details one from the
other. Likewise, the reading traditions preserved and transmitted
in each centre differed one from the other, mainly at the levels
of phonetics, phonology, and morphology (i.e., different dialects
of Hebrew), but also—to a far lesser degree—at the level of the
semantic content of the texts (Chiesa 1979, 9-36; Yeivin 1985,
21-36).

Each centre also developed its own distinct apparatus as an
aid to the preservation of minute details of the biblical text (con-
sonants + reading tradition)—its own masora, in the narrow sense
of the term. These textual commentaries differed one from the
other not only with respect to the texts they were designed to pre-
serve, but also in terms of the methods used, terminology em-
ployed, and even the way each commentary was preserved and
transmitted. The Tiberian Masora (t.Mas) exists as a vast nebula of

distinct notes and comments, which was drawn on according to

24 Phillips

need, taste, and availability, on a codex-by-codex basis (Yeivin
2003, 60-92). There is no authoritative standard text-form of the
t.Mas. By contrast, there is an authoritative standard text-form of
the Babylonian Masora (b.Mas), at least of the Pentateuch. The
b.Mas of the Pentateuch (b.MasP) was transmitted as a distinct
text, not usually ‘appended’ to the biblical text itself. Where sepa-
rate copies of the same portions of this text have been preserved,
they are functionally identical. Some Babylonian biblical MSS do
contain individual masoretic comments in the margins, but these
comments have their origin in a text with a fixed form (Ofer 2019,
151-57).!

In the same way that the Hebrew biblical commentaries of
R. David Qimhi and Abraham Ibn Ezra condemned the Judaeo-

Arabic commentaries of their forebears to obscurity for centuries,

' Abbreviations used throughout this article: A = Aleppo Codex; b.Mas
= The Babylonian Masora, referring to the entirety of the fixed text of
this textual commentary; b.MasP = The Babylonian Masora of the Pen-
tateuch (referring to the entire, perhaps unrecoverable, text of this tex-
tual commentary, rather than Ofer’s edition, which is fragmentary); DP
= Damascus Pentateuch; L = Firkowich B 19a, i.e., the ‘Leningrad’ Co-
dex; L” = Pentateuch MS, written by Samuel b. Jacob (scribe of L),
containing a large proportion of Babylonian masoretic notes (the ma-
sora magna of the MS has been edited by Breuer 1992; also known as
Gottheil 14; see Gottheil 1905); Mm = Masora Magna; Mp = Masora
Parva; Of.b.MasP = Ofer’s (2001) edition of the Babylonian Masora to
the Pentateuch; Or. 4445 = ‘The London Pentateuch’, located in the
British Library; S' = Sassoon 1053; t.Mas = the Tiberian Masora: the
vast, nebulous, collection of individual notes designed to preserve the
Tiberian recension of the biblical text.

The Masoretic Notes in RNL EVR II B 80+ 25

so the rapid triumph of the Tiberian biblical text and reading tra-
dition among the various Jewish communities quickly led to the
demise and obscuring of the once-influential Babylonian text and
reading tradition—and the b.Mas with it (Yeivin 1985, 22-24).
However, since the uncovering of the Cairo Geniza, the study of
the Babylonian biblical tradition in general, and the study of the
b.Mas in particular, have flourished.

The study of the masoretic notes from the Babylonian tra-
dition is very much a work in progress. Over the course of the
last century many individual fragments of b.Mas have been ed-
ited. This labour culminated in Ofer’s 600-page annotated edition
of the Babylonian Masora of the Pentateuch (Ofer 2001). This
edition incorporates all the currently known fragments of MSS
containing the b.MasP text itself (though not the individual notes
found in various biblical MSS). Nevertheless, he estimates that
his edition comprises only about one sixth of the original b.MasP
text. There is, to date, no edition of the b.Mas of the Prophets and
Writings, though Ofer is at work on the latter (see Ofer 2011, 148
n. 42).? For obvious reasons, the present article will make con-
stant reference to Ofer’s edition of the b.MasP.

Part of the difficulty in studying the b.Mas is that the ma-
jority of the relevant manuscripts are from the Cairo Geniza, with
all the fragmentariness involved with documents from this
source. However, a crucially important exception to this is the

manuscript known as L”. L™ is a Tiberian Torah MS in terms of

? Weil (1963) and Yeivin (1982) have each edited small portions of the
b.Mas of the Prophets, but compared to the b.MasP, the manuscript re-
mains are few and far between.

26 Phillips

text and layout, yet a very high proportion of its Mm seems to
have come from a Babylonian source. Breuer (1992) has edited
and annotated the Mm of L™ in two volumes, but the MS itself
remains inaccessible to scholars.* Ofer (2001, 13-25) devotes an
entire chapter to the discussion of L”, and uses it throughout
b.MasP.Ofer as a supplementary source to help clarify difficult
portions of the MSS containing the formal b.MasP text.

L™ is not the only Tiberian MS that makes use of the b.Mas
(or at least Babylonian graphemes and masoretic terminology).
Indeed, Yeivin (1968, 72-75), Breuer (1992), Ofer (2001, 260-
74), and Dotan (2005) have demonstrated that most of the well-
known Tiberian MSS contain a certain amount of masoretic ma-
terial originating from a Babylonian source. At one end of the
spectrum, this Babylonian influence is very slight (e.g., in A); in
other MSS (e.g., BL Or. 4445) the influence is more pronounced;
and in still other MSS the influence is rather substantial (e.g., S').
This phenomenon raises further questions, therefore, regarding
the nature of the interaction between, and mutual influence, of
the Tiberian and Babylonian masoretic traditions. At any rate, to
date, L™ is thought to stand alone in terms of the massive extent
to which the masran behind a Tiberian codex made use of Baby-

lonian material in his masora.*

3 For the purposes of the present article this inaccessibility is a particu-
lar frustration, inasmuch as the Mp notes of the MS are highly relevant
to our topic here.

4 The question of what ‘counts’ as a Babylonian masoretic note can be
answered in various ways. Breuer (1992) adopts a minimalist approach,
and classifies Mm notes in L™ as Babylonian only if he can demonstrate
that the content of the note does not match the Tiberian biblical text,

The Masoretic Notes in RNL EVR II B 80+ 27

However, it now appears that another Tiberian MS must be
placed close to L™ with regard to the extent of the Babylonian
influence on the masoretic notes: MS RNL EVR II B 80 +.° Though
quintessentially Tiberian in text-form and layout, the masoretic
notes of this Torah MS show very extensive and sustained influ-

ence of the b.Mas.°®

but does match what is known about the Babylonian biblical text and
dialect (Ofer 1992, 272-73). There are at least two obvious difficulties
with this approach. First, a great deal of the Babylonian biblical text
and dialect is identical with the Tiberian text and dialect. Therefore,
one would expect that a large proportion of the Babylonian masoretic
notes would be identical in content (even if different in form and ter-
minology) to the Tiberian notes. Breuer’s approach makes no allowance
for this, and therefore can only be expected to catch a relatively small
subset of the Mm notes in L™ taken from a Babylonian source. Secondly,
much about the Babylonian consonantal text (and, to some extent, dia-
lect) remains uncertain, and so, once again, Breuer’s approach can only
positively identify a subset of the genuinely Babylonian notes. Yeivin
(1968) is similarly cautious about identifying Babylonian notes in Tibe-
rian codices purely based on the appearance of Babylonian vowel signs
and masoretic terms, such as '>w or even 'p7. However, by comparing
the Mm notes in L™ with the content of his edition of b.Mas.P, Ofer
found that the great majority of the notes in L™ derived from the
b.Mas.P, even when the content of the notes matched the Tiberian text
just as well as the Babylonian text. In turn, this fully justifies taking a
broader approach to identifying Babylonian notes. This broader ap-
proach (relying on not only the content of the notes, but also their form)
is adopted by Ofer (2001) and Dotan (2005), and is the approach fol-
lowed in this article.

° The + sign indicates additional shelfmarks (see 81.0, below).

° To the best of my knowledge, the masoretic notes of this MS have
drawn scholarly attention twice before. Strack (1897) mentions two

28 Phillips

The purpose of this article is to present a preliminary report
on the masoretic notes (Mm and Mp) of this MS. The report will
demonstrate the nature and extent of the Babylonian masoretic
material therein, and begin to illustrate some of the ways in
which this material can be used to consolidate and expand our
existing knowledge of the b.Mas. Thus, the article is selective and
illustrative, and makes no attempt at comprehensiveness. It is
hoped that it will alert other students of the b.Mas to the riches
of this MS, until a full edition and analysis of its masora (already
underway) is completed.

After a brief description of the codex, the discussion of the
masoretic notes is divided into three sections: structural features,

external features, and internal features.

e Structural features: The b.Mas has a set of distinctive pat-
terns, or formats, in the way its notes are structured. Some
of these are scarcely attested in the t.Mas, but common-

place in the b.Mas. In the first section of this study, I show

notes (one Mm, one Mp) from the MS (which he refers to by its old
numbering: Cod. Tschuf. 51), both of which discuss variant readings
between the Easterners (Sura and Nehardea) and the Westerners. I am
grateful to Prof. Yosef Ofer for bringing this article to my attention. The
second discussion of the masoretic notes of this MS appears in Pen-
kower’s (2020) recent study of the textual variants between Sura and
Nehardea. The frequent Mm and Mp notes discussing textual variants
between the Eastern schools of Sura and Nehardea is further evidence
of the Babylonian influence on the masora of this MS. Nonetheless, since
so many examples have already been adduced by Penkower in his study,
this aspect of the masoretic notes will not be considered further in this
initial report.

The Masoretic Notes in RNL EVR II B 80+ 29

that almost all the quintessentially Babylonian note-for-
mats are found with great frequency throughout the MS.’

e External features: there are other (non-structural) Babylo-
nian aspects that are visible ‘on the surface’ of the notes
(i.e., self-evident to the eye, without having to analyse the
content of the notes). These include terminology from the
b.Mas, aspects of Babylonian Aramaic in the notes, etc.
These features are reviewed in the second section of this
chapter.

e Internal features: the final section of this chapter shows
that when the content of many of the notes is analysed, it
often only matches the Babylonian biblical text, or the Bab-
ylonian dialect of Hebrew.

1.0. Description of RNL EVR II B 80+8

The codex here referred to as RNL EVR II B 80+ is currently pre-
served under at least three shelfmarks. EVR II B 80 preserves the
lion’s share of the remains of the codex: 124 leaves.? EVR II B 170

” Ofer’s (2001) monumental study of the b.MasP has been immensely
instructive throughout this whole study. In particular, this section on
the various common formats of Babylonian masoretic notes relies al-
most exclusively on his descriptions and analyses of the various note
types.

8 A full description of this MS (rather than simply its masora) is currently
in preparation as a separate article. The brief description here, there-
fore, is intended only to orient the reader regarding the most general
aspects of the MS, before focussing on the masoretic notes.

° In its current presentation on the Ktiv website, as well as in the de-
scriptions in Beit-Arié et al. (1997, 13-14) and Dukan (2006, 310-11),

30 Phillips

and EVR II B 14 each contain an additional three leaves. Thus, 130
leaves of the original codex are currently available to us. Through-
out this article, individual shelfmarks are used where relevant, but
generally the entire codex is referred to, using the label EVR II B
80+.'° For ease, I will refer to the relevant image number when
giving examples from the codex. All the examples below happen
to be taken from EVR II B 80, which contains 256 images on the

Ktiv website.'! Thus, references take the form: ‘image x/256’.

the MS has 125 folios, and fol. 125 contains two colophons, including
a date (which has signs of tampering). However, it appears that this
folio is not, in fact, part of the same MS as the rest of RNL EVR II B
80+. The most obvious evidence for this comes from the fact that al-
though the double-dot sof-pasuq sign is used regularly and systematic-
ally throughout EVR II B 80+, it is virtually never used on fol. 125.
Additional evidence pointing towards the fact that fol. 125 is not from
our MS is to be found in the line-fillers employed, the positioning of the
masora circellus, the shape of the hatef-patah sign, and the use of rafe.
More details will be provided in the full description of the MS.

’° For the purposes of this study, I have searched through RNL EVR II B
1-600, looking for other fragments from this same MS. It is entirely
plausible that when the scope of the search is expanded (as it must be
before an edition of the notes can be produced), other portions of the
MS will be found. For this initial report, however, the 130 leaves found
thus far offer ample material for description.

" Accessible at https://web.nli.org.il/sites/nlis/he/manuscript.

The Masoretic Notes in RNL EVR II B 80+ 31

RNL EVR II B 80+ is a monumental (42 X 36cm), three-
column Model Torah Codex, probably from the 10th or 11th cen-
tury.!? The column height is 29cm, with 21 lines per column.'? The
margins are wide, with Mm in the upper and lower margins of
virtually all the pages (typically one or two lines of Mm in the
upper margin, and one to four lines in the lower margin). Mp notes
occupy the outer vertical margin, and the two inter-columnar mar-
gins, but rarely the gutter margin of the page. The Mm contains
some collative masoretic notes; these appear only in the upper
margin of the page and occasionally include select Tiberian signs
of cantillation and vocalisation (for example, the upper margin of
fol. 4r and again on 31r). The rest of the Mm and Mp is sporadi-
cally vocalised and/or cantillated, with Tiberian accent signs and
Tiberian or Babylonian vowel signs. These Babylonian vowel signs
are from the simple (rather than the compound) line system (see
Yeivin 1985, 54-55). Occasionally, the patah sign from the dot sys-
tem is employed (e.g., 55/256, 109/256). Many of the Babylonian
vowel signs are placed over the inter-consonantal space, a clear
marker of antiquity, according to Yeivin (1985, 55).

The biblical text is written in an accomplished Eastern
hand, with Tiberian vocalisation and cantillation signs. Conso-
nantally, the text is very close to that of A.

The extant portions of the MS comprise 130 folios. Of these,
the majority are well preserved, such that almost all the biblical

text on each leaf is extant and legible. Nonetheless, most of the

2 See n. 9 above regarding the dated colophon that has been hitherto
ascribed to this MS.

13 These dimensions are taken from Dukan (2006, 310-11).

32 Phillips

leaves contain some damage to the top margin and inner corner,
such that the upper Mm is frequently obliterated, partially or
completely, and often much of the Mp has also been lost. Fols
31-33 (images 65/256-70/256) and (especially) 46-49 (images
95/256-102/256) have suffered more extensive damage, such
that significant portions of the biblical text itself are no longer
legible on these leaves.

The following portions of the biblical text have been pre-
served in the MS (references in italics are from EVR II B 14; ref-
erences underlined are from EVR II B 170; other references are
from EVR II B 80):

Gen. 27.20-42.4; 42.4-33; 42.33-44.13; 46.10-50.26;

Exod. 1.1-2.3; 10.15-12.25; 13.2-14.28; 16.19-20.17;

21.28-30.38; 32.33-36.7; 39.15-42; Lev. 13.57-14.51;

14.51-15.24; 15.24-16.16; 17.6-18.21; 23.18-44; 25.8-

37; 26.42-27.23; Num. 1.23-51; 4.7-9.12; 10.20-13.23;

14.14-18.28; 20.1-28; 22.19-36.13; Deut. 1.1-31:14;
33.18-34.12.

In other words, almost all of Numbers and Deuteronomy have
been preserved, together with about two thirds of Exodus, half of
Genesis, and about a quarter of Leviticus.

Having been introduced to the MS itself, we can now focus
our attention on the masoretic notes therein. As mentioned
above, I will describe the Babylonian aspects of the notes from
three different angles: structural features (the typically Babylo-
nian ways many of the notes are constructed), external features
(other aspects of the notes that reveal their Babylonian origins,

without having to analyse the content of the notes), and internal

The Masoretic Notes in RNL EVR II B 80+ 33

features (the ways the notes reveal, when their content is ana-
lysed, that they refer to the Babylonian form of the biblical text
and the Babylonian dialect of Hebrew).

2.0. Structural Features

Ofer (2001, 75-123) gives an extremely helpful overview and de-
scription of the different types of masoretic notes that are typical
of the b.Mas. In this first section I simply demonstrate that almost
all of the quintessential forms of Babylonian notes are found, of-

ten in great profusion, in EVR II B 80+.

2.1. All-Inclusive Description”

One of the foundational differences between the Tiberian and
Babylonian masorot lies in the way each system counts and lists
whatever textual element is under discussion. The simplest way
to explain this difference is via an example.
(1)... ANT

‘And the men feared...’ (Gen. 43.18 [Image 42/256])
‘Igy

‘They feared’ occurs 6 times
pao ya "ow

4 Ofer (2001) refers to this as 55197 “ixvnn, and gives a comprehensive
discussion of the phenomenon. In Ofer (2019) this phrase is translated
in two ways: ‘general description’ and ‘all-inclusive description’. The
latter translation is more helpful, pointing towards the fact that the
count in these notes refers to the entirety of the Hebrew Bible, rather
than dividing the text into discrete sections and dealing with each sec-
tion in isolation, as is common in the t.Mas.

34 Phillips

plene in the Scriptures, thus:1°

All the Torah xmmisg 51
Josh. 10.2 mdi Py
1 Sam. 17.24 wen n& OMNI
2 Kgs 17.7 Sxqw a zon 20"
The Minor Prophets mvyy on
All the Writings, excluding: 10 pin-vans AyD
Neh. 6.16.1° oan 55 Nt

This note from EVR II B 80+ is concerned with the plene
spelling of wayyigqtol 1x7. Note in particular how the count of
six includes not only three individual verses, but also three large
stretches of text: ‘all the Torah’, ‘the Minor Prophets’, and ‘all the
Writings, except...’ By my count, the plene spelling of the way-
yigtol 1871 occurs six times in the Torah, four times in the Minor
Prophets, and twice in the Writings (Psalms). In this Babylonian
masoretic note, however, the entire Torah, all the Minor Proph-

ets, and all the Writings each increase the count by only one.

'S Translating masoretic notes into meaningful English is notoriously
difficult, due to the highly codified and condensed language in which
the notes are expressed. In the translations offered in this article I have
aimed at a lucid, idiomatic rendering of the sense of the notes, rather
than any sort of isomorphic formal equivalence.

'6 This note also appears in L™ at Gen. 20.8 (Breuer 1992, 124). There,
however, the language of the note has been ‘Tiberianised’, as has the
structure of the note, in such a way that the count no longer fits the
lemmata. In addition, the reference to ‘all the Torah’ has been omitted.
EVR II B 80+ thus preserves a more original Babylonian form of the
note, as well as its accuracy.

The Masoretic Notes in RNL EVR II B 80+ 35

There are here two distinctive features of the Babylonian All-In-
clusive Description: (i) the count is all-inclusive—all the Bible is
covered by the one count; (ii) entire books, or collections of
books, with multiple individual instances of the relevant textual
phenomenon, can be grouped together as ‘one’ instance for the

sake of the count.!’

'” By contrast, the typical Tiberian way of describing the same phenom-
enon would be to break down the biblical text into sections and deal
with those sections separately. For example, the note above could be
‘Tiberianised’ as follows:

‘They feared’ occurs three times plene, thus: payor "50 "34897
Josh. 10.2 mds yD
1 Sam. 17.24 WRT OX ONIN
2 Kgs 17.7 Sowa ison 2-7"
And all [occurrences in] the Torah, the xepnoi'wy oman do
Twelve, and the Writings are likewise 8 yA Tyan

plene, apart from one instance:
Neh. 6.16 oan 59 N77

Note how this Tiberian-style note divides the biblical text into two dis-
tinct sections. The first section (whose boundaries are typically unstated
and must be inferred from the note as a whole) consists of the Prophets,
apart from The Twelve. For this section, the count of plene occurrences
of ‘they feared’ is three (the clear minority—there are eight defective
forms in the same section). In the latter section: the Torah, the Twelve,
and the Writings, the situation is reversed: all the occurrences of ‘they
feared’ are plene, apart from Neh. 6.16. Thus, in the Tiberian-style note,
the biblical text is divided into suitable distinct sections, and each sec-
tion is dealt with separately. The note as a whole intertwines the dis-
cussion of the two sections using the term yinmiD7. For an overview of
the issues of counting and structure in masoretic notes (not specific to
the b.Mas), see Breuer (1976, 193-283).

36 Phillips

The Mm notes in EVR II B 80+ contain many dozens of ex-
amples of the All-Inclusive Description. Here is one more instance:
(2) APNIDN oa amg os Tay RY AVA ny" DR FAN...

...“Your servants are shepherds, as our fathers were.”
(Gen. 47.3 [Image 47/256])

Z q ! ! ! 'y 5
‘Our fathers’ occurs 11 times plene in ya "w 'x' maK

the Scriptures, thus: Dee!
Gen. 47.3 (end of verse) NPD} FIO WMALK D3
All of Joshua yuoin ain
All of Judges vay MIN
Jer, 3.24... mdax nwa
..and following (Jer. 3.25) ringdwi
Jer. 14.20 ™ YT
Jer. 16.19 mas oni
All of Psalms, excluding: jaa xndn wi
Ps. 22.5 IWnay inva 72
1 Chron. 12.1818 navyapmasx 7x xv
2 Chron. 29.9 imax bai 7m
2 Chron. 34.21. mar nw xd wWwRr dy

Once again, in this note, the count of 11 refers to the entire Bible,

rather than simply to one section thereof. Joshua, Judges and

18 Notice how the citations from Chronicles appear after the citations
from Psalms. This reflects the Babylonian arrangement of the biblical
books, rather than the Tiberian arrangement. See below, 83.4, for fur-
ther details.

The Masoretic Notes in RNL EVR II B 80+ 37

Psalms are each included in their entirety, even though they en-

compass multiple individual occurrences. !°

2.2. Partial Count Preceding the All-Inclusive

Description

Example (2) immediately above illustrates one of the systemic
dangers inherent in the All-Inclusive Description. Consider the
final five lines of the note:

jaa xndn wD

IWNax inva Ja

NAVIPMAX OX XP

mas bal nim

max Mw xo Awe dp
When one deciphers the biblical references behind the lem-
mata, it becomes clear that only one exceptional verse from the
book of Psalms is mentioned. The subsequent three lemmata,
from Chronicles, are part of the overall count of eleven plene-
plene occurrences of irniax. However, if one does not carefully
locate the references behind the lemmata, these lines could easily
be misunderstood as presenting four exceptional verses from the

book of Psalms. The problem, in other words, lies in the fact that

'? Breuer (1976, 209) discusses the oddities of the equivalent Tiberian
form of this note. In a very important chapter, Ofer (2001, 75-100)
shows how our growing knowledge of the b.Mas has the capacity to
explain many such oddities outlined by Breuer. Neither Breuer (1976)
nor Ofer (2001) had access to this note in its Babylonian form. Now it
has come to light, it is clear that Ofer’s explanation perfectly fits in this
instance, too.

38 Phillips

there is no obvious boundary indicating the end of the list of ex-
ceptional verses, and the resumption of the list of verses making
up the primary content of the note.

Ofer (2001, 79-81) suggests that this inherent source of po-
tential confusion was the motivation behind an alternative form
of the All-Inclusive Description also found frequently in the
b.Mas: the Partial Count Preceding the All-Inclusive Description.
Once again, this alternative form is found frequently in the Mm
of EVR II B 80+. The example below, from EVR II B 80, is also
found in b.Mas.P.Of:

(3) ..ODIS8 niva xyvi ATA (Q) NMP (K) NTP 7D
‘These were the ones chosen from the congregation, the
chiefs of the tribes of their fathers...’ (Num. 1.16 [Image

111/256])

‘Their fathers’ appears twice xmga ‘5d '2 oniax
plene in the Torah, thus: pawyoi
Num. 1.16 ATyAORNTD TDN
Num. 17.18 Anan Fans ow mKi
[But] all [the rest of] the Torah Rmx 1D
1 Kgs 9.9 TANI
1 Kgs 14.22 NIP"
1 Kgs 17.41 Dm12 03
Ezra 10.16 gory 73
Neh. 7.617° "gana Panda x1

20 The term mANina ‘the latter’ indicates, in this instance, that that verse
in Nehemiah is being referred to, rather than the nearly identical verse
from the parallel list in Ezra 2.

The Masoretic Notes in RNL EVR II B 80+ 39

These are all the defectively spelled 'ya'on pon

occurrences.

This note could have been phrased as a normal All-Inclusive De-
scription: ‘Their fathers’ is spelled defectively six times in the
Scriptures: All the Torah apart from [two lemmata] + five more
lemmata from Kings and Ezra-Nehemiah. The problem with this,
as described above, is the potential confusion in moving immedi-
ately from the two exceptional plene spellings in the Torah, back
to the five defective spellings in Kings and Ezra-Nehemiah. In-
stead, the Partial Count Preceding the All-Inclusive Description
relocates the two exceptional verses from the Torah to the very
beginning of the note, then continues with the normal All-Inclu-
sive Description (minus the count) thereafter.

This note, in the same form, appears in b.Mas.P.Of (Ofer
2001, 495-96). The two notes are functionally identical, except
that: (i) "bw has been Tiberianised to 'n in our MS; (ii) pam
has been added before the initial two lemmata in our MS; and
(iii) the order of citations from Kings is canonical in our MS,

whereas the first two are inverted in b.Mas.P.Of.

2.3. Rule-Stating Notes”

The b.Mas pays considerable attention to noting spellings in the
Hebrew Bible that are uniform throughout the entire biblical

text.” These observations are then phrased in short notes

21 555 nya in Ofer’s (2001, 105-7) terminology.

?2 By contrast, the t.Mas pays far less attention to words with consistent
spellings, focusing instead on words whose spelling is variable.

40 Phillips

throughout the b.Mas text. The Mm of EVR II B 80+ contains
many dozens of such notes. Here are just a couple of examples,

neither of which appears in L™ or in Ofer’s edition:

(4) we? PHAX YT 87 NV TONY WN fTNy TIN Fav yw]
DIST MM Fra) ong dy Noes WTI

‘And he humbled you and let you hunger and fed you with
manna, which you did not know, nor did your fathers
know, in order to make you know that man does not live

by bread alone...’ (Deut. 8.3 [Image 202/256])

The forms JyT1n and Jy syd yin
are always written plene. hyy "O15
(5) ...oyTN"nS MD? 82 ApS 73g? WD np pws NP

‘Not because of your righteousness or the uprightness of
your heart are you going in to possess their land...’ (Deut.
9.5 [Image 204/256])

Every occurrence of the forms 1w°n1, Ww ww wwe
sw’, and 1v is spelled defectively. 'on "1D

In addition to these Rule-Stating Notes in the Mm, the masran
behind EVR II B 80+ has also reworked many such observations
from the b.Mas into masora parva notes.”? For example, the fol-
lowing Mp notes all appear on image 79/256 (Exod. 25). Each of
them appears to have been taken directly from the b.Mas, which
is largely extant at this point:

23 In Dotan’s (2005, 36) overview of the traces of the b.Mas extant in
Or. 4445, he mentions the existence of many Mp notes of the format:
[12 'na/'on/"9n] + dia. He sees these notes as one of the most prominent
aspects of the Babylonian ‘residue’ in the masora of the MS.

The Masoretic Notes in RNL EVR II B 80+ 41

Table 1: Mp notes in EVR II B 80

Reference Biblical Mp note in Equivalent note in Babylonian
Text EVR II B 80 Masora (Ofer 2001, 450-51)

Exod. 25.12 rnava ‘on "1D ID yAYD DINYAT NYA rnnya
‘on

Exod. 25.17 maa ‘on di ‘on "to mam mad

Exod. 25.19 ana 24 "1D w "2 an29 an 113

Exod. 25.19 “100 on "19 w "12 a199 391 113

Exod. 25.20 man ‘on "iD ‘on "19 madA MAD

To iterate the point: the t.Mas (magna and parva) does not
tend to focus on uniform spellings, but variable spellings. None
of the notes above appear in L (at all), or in Or. 4445 or DP (ad
loc.). In fact, Yeivin (1968, 74) finds only one 'on "12 note (which
he, too, recognises as Babylonian in character) in the whole of A,
and no 72 'n>/'9w/"5n "1D notes at all. The masran of EVR II B 80+,
though creating a genuinely Tiberian MS with typical Tiberian
format for the masora (i.e., the distinction between Mm and Mp),
nonetheless populates the Tiberian MS with many notes from the

Babylonian masoretic tradition.”

4 Here, the masran has semi-Tiberianised the note by converting w to
"on. Nonetheless, the note remains quintessentially Babylonian, in that
it attends to a uniform, rather than a variable, spelling.

5 These Rule-Stating Notes are certainly not the only Babylonian notes
to be found in the Mp of EVR II B 80+. At present (until the MS is made
available for scholarly examination) we do not have access to the Mp
notes of L™, and thus cannot tell to what extent reworked Babylonian
masoretic content is to be found therein. Until access to L” becomes
available, EVR II B 80 + appears to be the most significant MS currently

42 Phillips

2.4. Rule-Stating Notes, with Exceptions

If a particular word has a uniform spelling apart from one or two
exceptions, the b.Mas has a typical formula for describing both
the regular spelling and the exceptions:

'ya bw/'on [ean pia] [-2 pin/7a 12/-] 'on/"ow "ia [Ada]

[the word] all plene/defective [except from] [select

verses] defective/plene...”°
First, the majority spelling is noted, in the typical format of the
Rule-Stating Note. Thereafter, the exception is noted. According
to Ofer (2001, 106), the two parts of the note are not usually
linked by any sort of prepositional phrase, e.g., } 32 or - pin,
and one is left to infer that the latter citation is an exception to
the previously stated rule, by means of the concluding clause:
'ya '5w/'on. In Ofer’s list, only 5 out of the 14 notes contain a
linking prepositional phrase.

EVR II B 80+ contains a great many Rule-Stating Notes,
with Exceptions. The evidence of the initial survey suggests that,
unlike in the pure b.Mas, the two parts of the note are usually
joined with the prepositional phrase 7m 72.7” Nonetheless, some

maintain the pure form dominant in Ofer’s edition, such as the

available for examining the process of embedding Babylonian masoretic
content into Tiberian-style Mp notes. I hope to carry out a full study of
this phenomenon soon.

6 This is a slightly emended citation of Ofer’s (2001, 106) formulation.

?” Possibly, this is evidence of an attempt to render the unfamiliar Bab-
ylonian form more readily understandable to a Tiberian user of the MS.

The Masoretic Notes in RNL EVR II B 80+ 43

following (this particular example is found in neither
b.Mas.P.Ofer nor L™):

(6) TT PPA
‘You shall prepare yourself a way...’ (Deut. 19.3 [Image

225/256])

The forms pan and p> are always ho pp pon
plene [except for:] "on
Ps. 89.3, Dna TNX yan ow
which is the only defective spelling 'ya'on

of jan in the Scriptures.”®

2.5. Classifying Notes”

If a particular word in the biblical text has two optional matres
lectionis, this results in four possible spellings for that word. The
b.Mas has a distinctive type of note for such forms. First, the var-
ious minority spellings are grouped, and their references listed.
Then, the note ends with the formula ‘and the rest are spelled +
[most frequent spelling]’.*° Thus, the note serves as a guide to all
the various spellings of that particular word in all its occurrences.
Ofer (2001, 108) notes that in the formal b.Mas text these notes
are usually introduced with the formula: 'om "bw + [nbn],

8 Once again, this note shows a token effort at Tiberianisation. None-
theless, the content and structure of the note, and the use of the term
'ya are all quintessentially Babylonian.

29 pamnAn monn niyn in Ofer’s (2001, 108-10) terminology.

3° Note that, because of the b.Mas’s distinctive way of counting (see
§2.1, above), the ‘minority’ and ‘majority’ spellings may well not be
classified on a purely numerical basis.

44 Phillips

whereas in L™ this formula has usually been removed (perhaps in
an effort to Tiberianise the notes).

EVR II B 80+ contains dozens of these Classifying Notes.
In almost all cases it seems that the introductory formula has
been removed (as with L™), though other typically Babylonian
terminology within the note itself often remains. The example
below (the only example I have thus far found with the introduc-
tory formula still extant) also appears in Of.b.MasP and L”,
thereby offering an excellent opportunity to compare the three

sources:

(7) ...732 Thaa DST ng yey) Wn py IaTa WyaTny

‘Bezer in the wilderness on the tableland for the Reubenites, Ra-
moth in Gilead for the Gadites...’ (Deut. 4.43 [Image 195/256])

The plene and defective spellings of nin: 'om Sw min
Deut. 4.43... 2772 182 1X
...and its parallel in Josh. [20.8] ywint ‘ait
Ezek. 27.16 3271 NNN
These occurrences are spelled nas: nD NOX
1 Sam. 30.27 nina wr
Ps. 18.28 nin Py
Prov. 6.17 nin ory
Ezra 10.29 nian dxwi
2 Chron. 18.19 on matt ba Sy
2 Chron. 22.5 onyya py
All of these are spelled mim: ‘np nia pnd
Job 28.18 wen NINA

Prov. 24.7 Sand mine

The Masoretic Notes in RNL EVR II B 80+ 45

(2 Chron. 18.19?)*! and

3! Breuer also finds this lemma at this point in the note in L". However,
there it is written as nna’ and unpointed. Breuer (1992, 687) interprets
this as a second lemma from the same verse as the preceding lemma:
Prov. 24.7. He offers no explanation as to why two lemmata are given for
the same verse, nor why the rather indistinctive form nna’ would be used
as a lemma at all. In Ofer’s (2001, 552) version of the same note the
lemma nn» and the following lemma 72ww™ are both listed after the end
of the note, quite out of place. Ofer, too, interprets nn5” as referring to
Prov. 24.7, and suggests that the repetition of references to this verse
points to the coupling together of originally separate masoretic notes.

The present manuscript seems to have had a similar Vorlage to Ofer’s
and (particularly) Breuer’s MSS. However, the relevant word has been
read and interpreted differently. Consonantally, 4 is read at the end of
the word, rather than n. Additionally, the lemma in the present MS has
been vocalised as the equivalent of Tiberian ma’. This seems to refer to
either 1 Kings 22:20, or 2 Chronicles 18:19: aXnx-nx Ama) Ai WAN
yx niaqa 5&1 by (Kings; the Chronicles version adds Dx1w 75n after
axnx, and spells Ramoth with plene vay).

The evidence of the present MS, taken together with the textual wit-
nesses offered by Breuer and Ofer, offers another way to interpret the
presence of this puzzling lemma. First, note the unexpected spelling in
part of Ofer’s version of the note:

‘n> na...

wan nian

. x) mm

"ND NANI
The forms wan nian and bnx> nim are spelled without x, even though
they are part of the section of the list that concludes: 'n> nix. It is plau-
sible that a copyist, momentarily confused by the lemmata w*an nia and
Sux> nian, and perhaps also noting the reference to Chronicles immedi-
ately following, mistakenly added the lemma 7n»”, as a reference to 2

Chron. 18.19, in which ‘Ramoth’ is genuinely spelled nia.

46 Phillips

1 Chron. 6.58 aww
1 Chron. 6.65 Dan ATT 7
These occurrences are spelled minx. "nD MIAN
All the other occurrences are spelled nian. 'nd nh ‘RWI

The above examples illustrate my claim that a large pro-
portion of the Mm notes in EVR II B 80+ are quintessentially
Babylonian in their structure. Before moving on to discuss other
Babylonian facets of the masoretic notes in this MS, it is worth
pointing out that at least one typically Babylonian note-type is
seemingly absent from the masora of the MS: the Cross-Referenc-
ing Note.®? The reason for this lack appears to be the fact that the
masran behind our MS abandoned the b.Mas’ ‘principle of the first
occurrence’ in the process of fitting the Babylonian notes to his
Tiberian MS.*?

3.0. External Features

So far, the discussion has focussed on the structural aspects of the
masoretic notes in EVR II B 80+, i.e., how occurrences are
counted and how the masoretic information is structured in the
notes themselves. I hope to have demonstrated that almost all the
patterns and formats considered by Ofer to be characteristic of
the Babylonian Masora (as opposed to the Tiberian) are found in
the MS—often with great frequency. We now proceed to what I
call the External Babylonian Features. By this I mean non-struc-

tural aspects of the notes suggesting their Babylonian origin that

2 Regarding this type of note, see Ofer (2001, 60-74).
33 On the principle of first occurrence, see Ofer (2001, 26-29).

The Masoretic Notes in RNL EVR II B 80+ 47

are visible ‘on the surface’ of the note, that is, without any need
to analyse the structure of the note or what information it pro-
vides about the biblical text. Most obviously, this category in-
cludes the use of Babylonian masoretic terminology, some of
which occurs in great profusion throughout the MS. Also in-
cluded in this section are evidence of Babylonian, rather than
Palestinian, Aramaic features in the wording of the notes; evi-
dence of the Babylonian pronunciation tradition when lemmata
are vocalised; the Babylonian arrangement of the biblical books
in lists of lemmata; references to the parasot and pisqgot in the

masoretic notes.

3.1. Babylonian Terminology

The terms most distinctive of the b.Mas—'p7, 'pya, and ';w—are
found in great profusion in this MS, each occurring hundreds of
times, usually in the Mm, though to a far lesser extent also in the

Mp.** Samuel b. Jacob, in L™, attempted to remove as many of

34 The term 'p7 is an abbreviation of 1777; it clarifies that it is the reading
of a particular form that is of interest, rather than its written form. 'ya
is an abbreviation of xa>ya; it clarifies that the note pertains to the en-
tire Bible, rather than a subset thereof. The interpretation of these two
signs has a long and somewhat tortuous history (see, among others,
Kahle 1902, 15-18; 1913, 177-79; Yeivin 1966; 1973). Ofer (2001, 46-
53) provides a helpful overview of the use of these terms, and the his-
tory of scholarship pertaining thereto. These terms lack precise Tiberian
masoretic terminological equivalents. By contrast, the term %v, i.e.,
xnbw ‘plene’ is simply the Babylonian equivalent of the Tiberian term
"On, ie., xn ‘plene’.

Some scholars, particularly Yeivin (1968, 74-75) have argued that
the value of the term w as an indicator of the Babylonian origin of a

48 Phillips

these Babylonian masoretic terms as possible or to replace them
with their Tiberian equivalents. The scribe behind EVR II B 80+,
by contrast, was far less concerned about removing these terms.
Consequently, there are many notes whose Babylonian origin is
revealed by EVR II B 80+ for the first time, sometimes with sur-

prising ramifications. Here is a particularly fruitful example:

(8) 02377) DNA DANI "72? NWR TDS
‘How can I bear by myself the weight and burden of you
and your strife?’ (Deut. 1.12 [Image 183/256])

The forms Tn PR DX TPA PR TDK
are always spelled plene. hw "i
The word Jn is always spelled defectively. 'on Jn

Everything about this note is distinctively Babylonian, from
the terminology "bw "nn, to the fact that it concerns consistent,

rather than variable, spellings, to the use of Babylonian vowel

masoretic note is rather limited. Rather, he suggests that the term was
known to the Tiberian masoretes, and that at least some of them used
it simply as a synonym for the more usual ”. Dotan (2005, 35-36),
too, appears to downplay the probative value of the appearance of the
term in Or. 4445. Ofer (2001, 265-66) articulates an alternative inter-
pretation of the evidence. He suggests that the appearance of Babylo-
nian terms in an otherwise ‘pure’ Tiberian MS (such as A) could be seen
as residual evidence of large-scale borrowing of quondam Babylonian
masoretic notes into the t.Mas, where they were generally Tiberianised,
thus obliterating the evidence of their Eastern origins. In any case, in
EVR II B 80+ the point is moot. The frequent appearance of w must
be considered alongside the mass of other evidence of the Babylonian
nature of many of the notes. The argument is cumulative.

The Masoretic Notes in RNL EVR II B 80+ 49

signs, to the fact that it shows a characteristically Babylonian in-
terest in the n/n distinction (Ofer 2001, 285-97). It does not ap-

pear in b.Mas.P.Ofer or L”, but it does appear in a modified and

Tiberianised form in none other than the Aleppo Codex:

Figure 1: Snippet from Aleppo Codex, fol. 213v

The forms 7’n and 7°71 are spelled plene da Pm pn
whereas 7n is spelled defectively. onqm

Yeivin (1968, 73) denies that the presence of Babylonian
vocalisation in the Mm of A implies “any dependence on the
b.Mas on the part of A.” Ofer (2001, 267), however, considers
this particular note in A, and is more hesitant. Based on the in-
terest in the n/n distinction and the Babylonian vowel signs, he
suggests “it is possible... that the masoretic note was copied from
a Babylonian source.” Ofer could not be more definite in his
claim, since at that point there was no direct evidence of such a
note in the b.Mas. However, now that such a similar note has
been found in EVR II B 80 +, we can conclude with a greater level
of confidence that this note in A is indeed of Babylonian origin.
This, in turn, reopens the larger question of the extent of the in-

fluence of the b.Mas on the masora of A.*°

35 In fact, the context of the note on the page itself, in A, may provide
some further corroboration of the note’s Babylonian origins; the details
of this point, however, would drift too far beyond the scope of the cur-
rent paper.

50 Phillips

The terms '77, 'ya, and "w are the most frequently occur-
ring Babylonian terms in EVR II B 80+, but are by no means the
only such terms. Particularly probative are cases where there is
clear linguistic opposition between the Tiberian and Babylonian
masorot. The following four examples of Babylonian masoretic
terms all occur regularly in EVR II B 80+ (alongside their Ti-

berian counterparts):°°

© For lists of Babylonian masoretic terms, see Kahle (1902, 15-18),
Yeivin (2003, 93-95), and especially Ofer (2001, 39-59). Several of the
terms that seem to have started out as Babylonian have made their way
so thoroughly into the t.Mas that it is debatable whether they should be
considered as distinctively Babylonian at all. Ofer’s list, for example,
contains far more ‘Babylonian’ terms than does Yeivin’s.

The Masoretic Notes in RNL EVR II B 80+ 51

Table 2: B.mas terms in EVR II B 80+

Babylo- Tiberian Example of Babylonian Location of ex-
nian Term Equivalent term in EVR II B 80+ ample (text and
image)
minx>w maini/xina | wwan...mapia dw "1" Jmx Gen. 39.9
(Kahle TMX MORWA PanRdwi TMX (33/256)
1902, 86 .Panxdwi mainsdwi
n. 7)
mom nam Sw Tox AK TD Exod. 32.27
ton} ..."orT1 TAX nddSann (95/256)
OT ATI
yin. m2 TD DYIIN ..NTINA "3 Po? Deut. 25.3
21nd Dw (236/256)
NIVTUNN min See example immediately
above.
RINT — 7a.72 oaTA ADR Ap "a Deut. 1.1
axes onata abxi In (182/256)

"DT RINT WT

Additionally, scattered throughout the MS there are many
other distinctively Babylonian masoretic terms, albeit occurring
sporadically. Some examples: °23(7) ‘next to, adjacent to’
(243/256); 125—Babylonian name for the letter lamed (24/256);
xnxinna ‘location’—when a particular word can refer either to a
place, or to something else, this term is used to specify that the
form is being referred to in its sense as a location (45/256);
pwanwnt—the Babylonian equivalent of the Tiberian term sevirin
(76/256); 71 yyo7—another equivalent to sevirin (38/256).

To give some idea of the frequency with which one encoun-
ters Babylonian masoretic terminology in this MS, a random sam-
ple of 50 pages was taken, across the full extent of the MS. Three

pages contained no distinctively Babylonian terminology; 16

52 Phillips

pages contained between one and four Babylonian terms; 31
pages contained five or more terms from the b.Mas.

Before finishing this section on Babylonian masoretic ter-
minology in EVR II B 80+, it is worth noting which terms do not
appear in the MS.*” Most prominent among these are the Babylo-
nian terms for the vowels and accents, which appear to be totally
absent from the MS. Instead, the masran consistently uses the Ti-
berian names for vowels and accents, in both the Mm and Mp.

Here is a stark example:

(9) sana way POPE YD yep who TTIW TAB Tp...
‘,..until you are destroyed and quickly perish because of
the evil of your deeds, in that you have forsaken me.’

(Deut. 28.20b [Image 243/256])

Every time they occur next to p>>yn a3 519
the nouns ‘your (MS/PL)/their deeds’ om yn om>d5yn
the form of y7 has holem, yin yan yr
apart from a single counterexample: Jn yaa
1 Sam. 25.3, addyn yi nwp wx
where yn, uniquely, has patah. xnna.'pt

This note has many Babylonian distinctives. Structurally, it
is a Rule-Stating Note, with Exception (see 82.4, above). The
terms 713 and 'p7 are Babylonian, and the vowel signs are Baby-
lonian. Despite all this, the vowel at the end of the note is referred
to using the Tiberian (though also appearing in Babylonian)
xnna, rather than the more distinctively Babylonian nmi) nna".

3” Until an exhaustive analysis of the manuscript has been carried out,
the following observations are somewhat provisional.

The Masoretic Notes in RNL EVR II B 80+ 53

3.2. Babylonian Linguistic Elements

As with b.Mas.P (Ofer 2001, 46), the Aramaic of the masoretic
notes in EVR II B 80+ frequently contains phonetic and morpho-
logical elements typical to Babylonian Aramaic, rather than Jew-
ish Palestinian Aramaic. To give three illustrative examples: (1)
The Babylonian determined plural suffix abounds in the forms:
~ai ‘the Prophets’, e.g., 21/256, »2"n> ‘the Writings’, e.g., example
(1) above. Likewise, this suffix appears in some of the names of
the biblical books: waw ‘Judges, e.g., example (2) above, 725n
‘Kings’, e.g., 132/256. (2) ’Alef, rather than yod, is regularly used
in the gentilic suffix, e.g., [n]xatp (113/256), [n]xina
(132/256), axn>n (133/256). (3) There is an example of a Bab-
ylonian Aramaic aqtel infinitive (aqtole) on image 237/256:
These are all the occurrences of the form sor pon...

sor with the sense ‘increase, do again’. .ADIDINA 'DT

3.3. Vocalisation Reflecting Babylonian

Pronunciation

Closely related to these linguistic elements are the occasional vo-
calisations found in the Mm notes that reflect the Babylonian pro-
nunciation tradition of Hebrew. Many of the Mm notes in EVR II
B 80+ are vocalised (at least partially) with Babylonian vocali-
sation signs. This is not in and of itself a definitive mark of the
direct influence of the b.Mas, as most of the key early Tiberian

codices occasionally employ the Babylonian vocalisation signs in

54 Phillips

their Mm notes (Yeivin 1968, 72-74).*° Moreover, Yeivin (1968,
74) is careful to note that, although the graphemes used are occa-
sionally Babylonian, the dialect those graphemes represent is reg-
ularly Tiberian: “In all the MSS I have examined, the dialect re-
flected in the Babylonian graphemes is the Tiberian dialect.”

In EVR II B 80+, by contrast, an initial overview has re-
vealed at least two loci where the Babylonian vocalisation signs
reflect the Babylonian, rather than the Tiberian, pronunciation
tradition:

(10) TR WP nyna ...71wA NPN 1h

‘And he bought the piece of land... for a hundred qasita.’

(Gen. 33.19 [Image 21/256])

The form no’wp occurs three times, "no wp
thus: pAwyoi
Gen. 33.19 npn nx
Josh. 24.32 atwn npona
Job 42.11. AAT OF WRI

It appears that in the first lemma cited, a Babylonian hiriq
is marked over the het of npn. This accords well with the Baby-

lonian pronunciation tradition (see Yeivin 1985, 814), in which

38 Having said that, the extent of the use of Babylonian vocalisation
signs in the Mm (and, rarely, the Mp) of the MS is pronounced—some-
where between a third and a half of the notes containing vocalisation
signs employ Babylonian, rather than Tiberian, signs. (This count ex-
cludes the many collative masora notes, which are typically Tiberian,
and are always vocalised with Tiberian signs.) Such a high proportion
of Babylonian vowel sign usage does appear to point to the Babylonian
origins of much of the MS’s masora.

The Masoretic Notes in RNL EVR II B 80+ 55

the lowering of the original hiriq to a segol (as found in the Tibe-
rian tradition) is less operative, presumably due to the weakening

of the guttural het in the Babylonian pronunciation.”

(11) dating nfba-by DADNI
‘so that I may write on the tablets the words’ (Deut. 10.2
[Image 206/256])

The forms anax) and 3n31 ans anaNi
are always written defectively, except 7a na'on "nD
for Hos. 8.12, where 21n>x is plene. "Sw antin 27% ands

The biblical text itself at this point in the MS is written, as
expected, according to the Tiberian pronunciation tradition,
an2xi with the expected segol as the preformative vowel of the
1cs imperfect. However, the Babylonian vocalisation in the mas-
oretic note marks this preformative vowel as a hiriq, as is regu-
larly found in the early and middle forms of the Babylonian dia-
lect (Yeivin 1985, 449).

3.4. Babylonian Arrangement of the Biblical Books

The arrangement of the books of the Latter Prophets and Writings
differs somewhat between the Tiberian tradition and the Babylo-
nian.*° The most prominent differences pertain to the locations
of Isaiah and Chronicles (though there are various additional dif-

ferences regarding the order of the shorter books of the Writings).

3° On this weakening in Jewish Babylonian Aramaic, see Morgenstern
(2011, 73-76).

*° The order of the biblical books in the Babylonian tradition follows
the order found in the famous baraita in b. Baba Bathra 14b-15a.

56 Phillips

In the Tiberian tradition, the order of the Latter Prophets is
Isaiah, Jeremiah, Ezekiel, The Twelve; and Chronicles is located
at the beginning of the Writings.*' In the Babylonian tradition,
the order of the Latter Prophets is Jeremiah, Ezekiel, Isaiah, The
Twelve; and Chronicles is located at the end of the Writings.
These differences in sequencing crop up regularly in the
masoretic notes. In general, masoretic notes arrange their lem-
mata in canonical sequence.” Thus, if a particular note happens
to cite both Psalms and Chronicles—in that order—this is evi-
dence that the note has been drawn from the b.Mas.** Example 2
above illustrates this Psalms-Chronicles sequence of lemmata. Or
if, to give another example, a particular note cites from Jeremiah
or Ezekiel, followed by Isaiah, this is likewise evidence of Baby-
lonian origin. Such a note is found, for example, on image
133/256. After citing eight verses from Jeremiah, the note con-
tinues: ‘And all of Ezekiel, and all of Isaiah, and all of The

Twelve’.

“! This is the order of the books in the great Tiberian codices, such as L
and A. BHS locates Chronicles at the end of the Writings, and thus is
not a faithful reflection of L in this respect.

” This rather general statement has many exceptions—not least given
the ‘interwoven’ nature of many longer masoretic notes from the Tibe-
rian tradition, and the arrangement of the Babylonian Classifying Notes
(see §2.5, above).

#8 Breuer (1992, 12), followed by Ofer (2001, 16, 124-25), see this as
one of the most significant identifiers of notes from the b.Mas.

The Masoretic Notes in RNL EVR II B 80+ 57

3.5. Reference made to Pisqot/Parashiyyot and

Parashot

The b.Mas includes far more references to pisqot/parashiyyot and
parashot than the t.Mas.** They play a double role in the b.Mas:
sometimes rules are framed such that they pertain to just a single
pisqa or parasha; on other occasions they are used as reference
points to help identify the precise location of a given lemma (Ofer
2001, 126-34, 151-52). Both uses are found in the masoretic
notes of EVR II B 80+, more often in connection with pisqot,
though in isolated instances also with reference to parashot. In
the first example below, a note is formulated with reference to
just one pisqa. In the second example, a pisqa is used as a refer-
ence point to identify the particular occurrence of a given collo-
cation. In the third example the term xnwnp is used, referring to

Parashat Balaq.

(12) 733 mend) ..naya Ted ..onw cya Ten) Woe Men
maa mes) ..oney7a me) ...nvancaa mee
Soper "aa MDP .. ways M2) wwe TE?

‘of the tribe of Judah... and of the tribe of the sons of
Simeon... of the tribe of Benjamin... and of the tribe of

the sons of Dan... of the tribe of the sons of Menasseh...

* Ofer (2001, 151) suggests that the reason the parashiyyot play a far
larger role in the b.Mas than in the t.Mas is that the b.Mas specifically
attends to the preservation of the parashiyyot. At each relevant point in
the text, the b.Mas notes the presence of a new parasha, and the nature
of that parasha (petuha or setuma). By contrast, the t-Mas is seemingly
‘blind’ to the parashiyyot, not attending to the accurate preservation of
that aspect of the biblical text.

Phillips

and of the tribe of the sons of Ephraim... and of the tribe
of the sons of Zebulon... and of the tribe of the sons of
Issachar... and of the tribe of the sons of Asher... and of
the tribe of the sons of Naphtali... (Num. 34.19-29 [Im-
age 178/256])

In the pisqa of ‘These are ADRT Xpova mM 1D
the names of the men who WWE DWNT ninw
will divide for you’ (Num. 34.16-29) oad ony
conjunctive waw always precedes the monn

phrase ‘of the tribe of, except in three mond "37a 73

cases, as follows: pain
‘of the tribe of Judah’ (v. 19); anim mond
‘of the tribe of Benjamin’ (v. 21); jaa nen
‘of the tribe of the sons of Manasseh’ (v. 23). nwin 73a nvNd
These make the mnemonic 'p'a". qoro 'o'n"
Judah and Benjamin alone mot aa ata
among them omit ‘the sons of’. prs. yA
These make the mnemonic '12". sare ‘a9
Judah, Simeon, and Benjamin jaa pyaw AN
alone among them omit ‘a chief’. pia xwi nt
These make the mnemonic '1'w". qoro 'a'w"

The phraseology of the list of cis-Jordan tribal chiefs re-

sponsible for the distribution of the land inheritance is highly

stylised and repetitive. Nonetheless, there are many fine-grained

deviations. This long note, pertaining only to this single parasha,

The Masoretic Notes in RNL EVR II B 80+ 59

codifies these deviations to preserve the list from future modifi-

cations.*

(13) snpy mm yap wy Dvdyy boa MP2) SwAD MWD oem
‘And you shall become a horror, a byword, and a taunt
among all the peoples where the LorD will lead you away.’
(Deut. 28.37 [Image 244/256])

cf. ABI? priya nizban de> nit) (Q) mt (K) ayn bennn

niapardaa MPP?) Maw? owiy4

‘T will make them a horror to all the kingdoms of the earth,

to be a reproach and a byword, a taunt anda curse in
all the places where I shall drive them. (Jer. 24.9)

In the Torah, the phrase is: NOTIN
‘A horror, a byword, and a taunt’ Swnd nnw>
(Deut. 28.37). arw
The mnemonic for this is 'y'n'w. aro y'n'w
[But in the pisqa]

‘Like the [bad] figs’ (Jer. 24.8-10) DUNN
the phrase is: ‘And you will make them onnn
a reproach and a byword, Swndi naan
a taunt and a curse’ (v. 9). md>pn npw
The mnemonic for this is 'p'w'n's. aro 'p'w'n'p

* This note, in a similar form, also appears in L™ (Breuer 1992, 662).
However, in L™, the introductory phrase xpovp 7519 has been modified
to 'r3ay 52. Ofer (2001, 129-30) observes a similar tendency: notes in the
b.Mas mentioning weekly parashot are sometimes emended when taken
over into the t.Mas, with the reference to a parasha being replaced by a
more general reference to the ‘inyan.

60 Phillips

In this Comparative Masoretic Note,” the issue is the simi-
larity between some of the language of the covenant curses in
Deut. 28 and Jeremiah’s ‘fig-oracle’ (Jer. 24). The note compares
the wording of a particular phrase between the two texts, so as
to preserve the precise details of each, and prevent the one con-
taminating the other. The point for our purposes here is that the

reference is to the siman of the pisqa and not to the verse.

(14) ..005) soy omg TD) ea D> naw ...xit9 hn

wD? kT?

‘So now, come please... Now, come... And come please...

Come curse for me Jacob and come... Come please...
Come...’ (Num. 22.6-25.14 [Image 153/256])

[The lengthened imperative on 't 25]
na occurs 7 times in Parashat Balagq: [pao 'wI53
Num.] 22.6 xi no[> anyy]
Num. 22.11 Rx OYA An
Num. 22.17 JTIDKX T1979
Num. 23.7 twice in the verse pIoal OW INT OX jn
Num. 23.27 NPR
Num. 24.14 TRYN

“© Comparative Masoretic Notes are very common in the b.Mas, and
there are many such notes in EVR II B 80+. However, since they also
occur with some frequency in the t.Mas, this type of note has not been
adduced as particular evidence of the Babylonian quality of the masora
of EVR IIB 80+.

The Masoretic Notes in RNL EVR II B 80+ 61

The beginning of this partially obliterated Mm note has
been reconstructed on the basis of its associated Mp note, which
appears several times throughout the parasha: 'wnp3 "57 '.

There are many other instances of the lengthened impera-
tive n>» throughout the biblical text; the scope of this note, how-
ever, is limited to one weekly parasha.

This completes the overview of the external Babylonian
features of the masoretic notes in EVR II B 80+. Some of these
features, e.g., the use of Babylonian masoretic terminology, are
extremely common in the MS, while others, e.g., references to the
parashot, are rare. Taken together, however, these external Bab-

ylonian features colour almost every page of the MS.

4.0. Internal Features: Notes Reflecting the

Babylonian Textual Tradition

The discussion thus far has focused on the Babylonian elements
of the masoretic notes themselves. We have not yet ‘peered
through’ the notes, using them like a window, to examine the
consonantal text they aim to preserve, or the dialect of Hebrew
they presuppose. When the notes are interpreted in this manner,
they sometimes point to the Babylonian consonantal text and di-
alect, rather than the Tiberian.

An initial overview of the MS reveals that a number of the
masoretic notes reflect the Babylonian recension of the biblical
text, rather than the Tiberian. Many of these notes appear in the
Mp, rather than the Mm. As Ofer (2001, 264) has already sug-
gested, this is likely due to the fact that it is easier to ‘Tiberianise’

the Mm notes than the Mp notes, since the former include the

62 Phillips

relevant biblical references, and can thus be relatively easily
cross-checked to confirm that they match the Tiberian text. None-
theless, there are also some Mm notes in the MS that apparently
reflect the Babylonian recension of the MT. Space constraints
limit me to offering a single example each from the Mp and the
Mm. The first is already known to scholarship, whereas the sec-

ond appears to be hitherto unknown.

(15) OWS) anawa gn pwNLy MTN HR sam Ma) HXany 7A

2H 1a

‘Why did you flee secretly and trick me, and did not tell

me, so that I might have sent you away with mirth and

songs, with tambourine and lyre?’ (Gen. 31.27 [Image
16/256])

The Mp note related to nwz1 ‘and with songs’ reads: 'w51 'on 'n
‘This word, and those like it, occurs 5 times written defectively
with this vocalisation, i.e., a hireq without a mater lectionis’.
Breuer (1992, 15, 183) has already discussed this note,
which appears in the Mm of L™. The count of five seems to match
only the Babylonian text, rather than the Tiberian. The equiva-
lent Tiberian note counts only four such defective forms. The ad-
ditional occurrence in the Babylonian version of the note occurs
at Jdg. 5.1. The Tiberian text reads 79127 -Wmi ‘and Deborah
sang’, but it seems that the Babylonian text must have read wm

with hirig.””

*” As Breuer himself notes (ad loc), Yeivin (1985, 654) catalogues an
equivalent phenomenon with the root 1"10 in the Babylonian tradition.

The Masoretic Notes in RNL EVR II B 80+ 63

(16) ...cAdp OOP wea inariza Vaya 72 RYAN?
‘There shall not be found among you anyone who burns his
son or his daughter as an offering, anyone who practices
divination...’ (Deut. 18.10 [Image 223/256])

Forms of the verb nop are always spelled ‘DD DOP
defectively, [apart from] on
Ezek. 17.9 DID WD NI
Zech. 10.2, Wm OOP
the only plene forms in all the Scriptures. 'ya "ow

This intriguing Mm note appears to preserve a different con-
sonantal text of Ezek. 17.9 to that known from the Tiberian tradi-
tion. First, observe that the formulation of the note is quintessen-
tially Babylonian: a Rule-Stating Note, with Exceptions (see §2.4,
above). Likewise, fully Babylonian are the terminology, "n>, ">w,
'ya, and the vocalisation signs. There is little doubt, therefore, that
this formulation is a genuine part of the b.Mas. The note is con-
cerned with the root o"vp. It claims that words formed from this
root are always spelled defectively, save two exceptions, which are
then listed. The first exception is from Ezek. 17.9. In the Tiberian
text the relevant clause reads: way dpip" | A A|"NNI ‘and its fruit cut
off. The Babylonian text, by contrast, apparently read nip’. This
cannot simply be a lapsus calami on the part of the masran, since
the inclusion of Ezek. 17.9 in a note dealing with the root o"op
makes sense only on the basis of the reading odip rather than
ooip’.*® Thus, this consonantal difference should be added to the

‘8 The masran of EVR II B 80+ seems to have struggled with this word
when writing the note: a samekh and final mem are visible, superim-
posed, at the end of nip’. The image quality is not sufficiently high to

64 Phillips

growing list of textual differences between the Babylonian and Ti-
berian recensions of the biblical text.*

These examples (and many others like them) notwithstand-
ing, it must be noted that the great majority of the Mm and Mp
notes in the MS are consistent with the Tiberian text and dia-
lect—including those notes with distinctively Babylonian struc-
tural and external features. This may simply be because the Bab-
ylonian and Tiberian recensions of the biblical text share a large
amount of material. Alternatively, it might be that the masran,
while content to keep the Babylonian form and language of the
notes, was careful to try to use only such notes as are consistent
with the Tiberian text. It is also possible that the masran kept the
Babylonian form and language of the notes, but actively altered
the content of the notes, when necessary, to match the Tiberian

text and dialect. For example, consider the following note:

determine which letter was written first, and which is the supposed cor-
rection.

* T am very grateful to Prof. Yosef Ofer, who (after a lecture in which I
presented this Mm note) checked the three known Babylonian biblical
MSS containing this verse and confirmed that one of them (Oxford Bod.
Heb. d.49 18v) does indeed read 0017”. This consonantal variant will be
considered from a text-critical angle elsewhere.

The Masoretic Notes in RNL EVR II B 80+ 65

(17) ...0Ax? NID nya cy DWT nx
‘And I commanded Joshua at that time...’ (Deut. 3.21 [Im-
age 190/256])

The Westerners spell ‘Joshua’ plene 'ynd> Sn 'a pwin
twice, as follows: armeTehen|
Deut. 3.21 NY pIWIT? NONI
Josh. 24.31 ~waWT ON ITA
(the second occurrence in the verse). NpIDaT NPN

The same note appears in b.Mas.P.Ofer and L”, except that
in these sources the count is given as three, rather than two, and
the final line of the note reads 'na '1w (Ofer 2001, 541) or its
equivalent, 'p1051 '1 (Breuer 1992, 679). In Ginsburg (1880, 213)
the beginning of the note reads 'm01 "5n '2 ywin’, omitting the
reference to the Westerners. Obviously, one cannot form gener-
alisations based on such slight evidence, but it is plausible to in-
terpret the note here in EVR II B 80+ as a self-conscious rework-
ing of the Babylonian note in light of the Tiberian text. Hopefully,
the full analysis of the masora of this MS will enable more clarity

on this point.

5.0. Conclusion

In this article I have attempted to show that the masoretic notes
(Mm and Mp) of RNL EVR II B 80+ have been deeply influenced
by the Babylonian masoretic tradition—perhaps more so than in
any other Tiberian MS (apart from L™) of which we are currently
aware. In fact, the extent of this influence is so pronounced, that

there is scarcely a page which does not reveal at least some trace

66 Phillips

of Babylonian Masora. The Babylonian nature of the notes is ap-
parent in the structure of the notes, the language and terminol-
ogy employed, and sometimes the content of the notes—i.e.,
what the notes claim about the biblical text. On the basis of an
initial sample of 35 Mm notes from Num. 31-34, about half of
the notes were also found in b.Mas.P.Ofer. Moreover, about two-
thirds of all the notes were either found in b.Mas.P.Ofer or
showed clear external signs of having been drawn from the
b.Mas. This survey ignored Mp notes, some of which also showed
clear Babylonian traces.

Despite the extent of the Babylonian influence, the masran
responsible for the codex was apparently very accomplished at
making the masora compatible with the Tiberian nature of the
MS. This is perhaps most obvious in the apparently total absence
of Babylonian names for the vowels and accents. There is also
some evidence that the masran may have moved beyond simply
choosing Babylonian notes compatible with the Tiberian text, to
emending Babylonian notes such that they fit the Tiberian text
rather than the Babylonian text, even though their style and ter-
minology remain Babylonian. Nevertheless, this process of filter-
ing and possible emending was not accomplished perfectly, and
a small but important proportion of these Mm and Mp notes re-
main incompatible with the Tiberian text. These notes provide a
window into the Babylonian text and reading tradition.

The nature of the codex—tTiberian in text and mise-en-page,
but with Babylonian masoretic notes—invites particular compar-

isons between this MS and L™. Pending fuller examination of the

The Masoretic Notes in RNL EVR II B 80+ 67

MS, all such comparisons remain provisional. Nonetheless, some

initial observations can be made:

(a)

(b)

(c)

The present MS retains the quintessential Babylonian
terms 'p7, w, and 'p2 to a far greater degree than L™. This
may allow us to identify a Babylonian background to
many more masoretic notes than has hitherto been
possible.

The present MS contains many Mp notes influenced by
the b.Mas, and the question of the reworking of the b.Mas
into Tiberian-style Mp notes requires significant further
study. This reconfirms the desideratum of gaining access
to digital images of L™, such that the Mp notes therein can
be similarly studied.

L™ is well-known for the many errors in its masoretic
notes (Breuer 1992, 12-15). The evidence thus far
suggests that the notes in EVR II B 80+ have been far
more carefully and competently copied. Several of the
examples above have already shown that the present MS
has the potential to resolve some of the outstanding
difficulties in the notes of L™, and even in Ofer’s edition
of the b.Mas itself.

I will finish with one final example where the present MS

resolves a minor difficulty in b.Mas.P.Ofer by filling in a series of

lacunae.

(18) :n77¥ 9230 NNSIN 1 npn NI? NNN 97 790

‘From Mount Hor you shall draw a line to Lebo-hamath,
and the limit of the border shall be at Zedad.’ (Num. 34.8

[Image 177/256])

68 Phillips

There are two occurrences of mixyin nxeyin 'noT nixgyin
spelled plene-defective: pawyoi '3
Num. 34.8 won tna ajnn
Ezek. 48.30. syyr meyin 7d
Prov. 4.23 is the only occurrence Ow mxxin nn 72
spelled plene with two vavin. Pana ‘ya
All the other occurrences are written mayn map ‘xwi
Minyn. nD

This same note is found, with three substantial lacunae, in
EVR II B 1549, the largest of the fragments of b.Mas that Ofer
uses for his edition, and which was previously edited by Ginsburg
(1885). These lacunae render possible widely differing recon-
structions of the note, as comparison between Ginsburg’s and
Ofer’s editions shows:
[ndan ma no njxyin yn nexin mdx1 nan ann [om dw nxexijn
n> in[xyn pan] na ya bw

(Ginsburg 1885, 242)

The citation of 753n na refers to Josh. 18.19, and the cita-
tion of 7>n nr refers to Josh. 16.3. The reconstruction of the third
lacuna is particularly unconvincing: the two letters preceding the
lacuna are clearly n1 rather than m1, and after the lacuna the
letters n> n (without a vay) are certain. Similarly, at the end of
the first lacuna a n is clearly visible, which Ginsburg ignores.

[oyn no n]xxin yn nxyin adxi ann ann nl? om ow mixin
nD Nk LIXWi nD nXyn PPaATIT NINA ya bw

(Ofer 2001, 518)

The Masoretic Notes in RNL EVR II B 80+ 69

Ofer expresses hesitations regarding his proposed recon-
struction of this note. The words 77177 X1n1 are a reference to
Josh. 15.11, where Ofer found this defective-defective reading in
a Babylonian MS. However, reconstructing the three lacunae ac-
cording to Ofer’s proposal reveals that his hesitations were well-
founded. Here is a reconstruction of the first lacuna in EVR II B

1549 according to Ofer’s proposal:°°

ornbweorny tp:

As the image shows, there is plenty of space to fit Ofer’s

proposed reading between the tav at the beginning of the line and
the tav at the end. However, the reading leaves a small amount
of space at the end of the lacuna, and it is not at all obvious what
might fill that space and connect to the tav immediately after the
lacuna.

Ofer’s proposed reconstruction of the second lacuna fits the
available space, narrowly, but his proposal regarding the third

lacuna does not:

jor” ys

WESTIN FONSI TH yi
In the reconstruction above, I have abbreviated Ofer’s xin.

to ana, but even so, the text is substantially too long for the avail-

able space.

5° All the reconstructions below were created using combinations of let-
ters and words from the same side of the same leaf of EVR II B 1549 in
which the lacunae are found.

70 Phillips

Based on the note in EVR II B 80+, emended in light of the
typical phraseology and scribal tendencies in EVR II B 1549, I

suggest the following reconstruction:

‘ ‘ oy ‘
Sworn mixyin dyn nein qdsxi an ann n[xin not nixeijn
n> n[igyn xAxXwiIn pajna ya

The proposed reconstruction of the first lacuna fits the
available space well:

WVHA INTO VIVID

The proposal for the second lacuna is plausible, but does

COT URS

leave a small amount of free space:

FOSHAN Rien 444 YS

The reconstruction of the third lacuna fits very well:

If this reconstruction proves persuasive, then this note from
EVR II B 80+ has solved one small mystery regarding the text of
the b.Mas, and also gone some way to clarifying the Babylonian
biblical text itself. It is to be hoped that a full edition and exami-
nation of the masoretic notes in this important MS, already un-
derway, will yield a great many further insights of a similar na-

ture.

The Masoretic Notes in RNL EVR II B 80+ 71

References

Beit-Arié, Malachi, Colette Sirat, and Mordechai Glatzer. 1997.
Codices hebraicis litteris exarati quo tempore scripti fuerint ex-
hibentes. Vol. 1. 4 vols. Monumenta Palaeographica Medii
Aevi, Series hebraica 1. Belgium: Brepols.

Breuer, Mordechai. 1976. The Aleppo Codex and the Accepted Text
of the Bible. Jerusalem: Mosad Harav Kook. [Hebrew]

(ed.). 1992. The Masorah Magna to the Pentateuch by

Shemuel ben Ya‘agov. 2 vols. The Manfred and Anne Leh-

mann Foundation Series 16. New York: Manfred and Anne
Lehmann Foundation. [Hebrew]

Chiesa, Bruno. 1979. Emergence of Hebrew Biblical Pointing: The
Indirect Sources. Frankfurt: Peter Lang.

Dotan, Aron. 2005. ‘Babylonian Residues in the London Penta-
teuch Codex’. In Studies in Bible and Exegesis 7, edited by
Shmuel Vargon, Yosef Ofer, Jordan S. Penkower, and Jacob
Klein, 33-40. Ramat Gan: Bar-Ilan University Press. [He-
brew]

Dukan, Michéle. 2006. La Bible hébraique: Les codices copiés en
Orient et dans la zone séfarade avant 1280. Bibliologia: Ele-
menta Ad Librorum Studia Pertinentia 22. Belgium:
Brepols.

Ginsburg, Christian D. 1880. The Massorah Compiled from Manu-
scripts: Alphabetically and Lexically Arranged. 4 vols. Lon-
don. [Hebrew]

Gottheil, Richard. 1905. ‘Some Hebrew Manuscripts in Cairo’.
Jewish Quarterly Review 17/4: 609-55.

72 Phillips

Kahle, Paul. 1902. Der Masoretische Text des Alten Testaments nach
der Uberlieferung der Babylonischen Juden. Leipzig: J.C. Hin-
richs’sche Buchhandlung.

. 1913. Masoreten des Ostens: Die Altesten Punktierten Hand-

schriften des Alten Testaments und der Targume. Beitrage zur
Wissenschaft vom Alten Testament 15. Leipzig: J.C. Hin-
richs’sche Buchhandlung.

Khan, Geoffrey. 2020. The Tiberian Pronunciation Tradition of Bib-
lical Hebrew. 2 vols. Cambridge: Open Book Publishers and
the Faculty of Asian and Middle Eastern Studies.

Morgenstern, Matthew. 2011. Studies in Jewish Babylonian Ara-
maic: Based on Early Eastern Manuscripts. Harvard Semitic
Studies 62. Winona Lake, IN: Eisenbrauns.

Ofer, Yosef. 1992. ‘The Babylonian Masorah to the Pentateuch in

a Tiberian Recension’. Leshonenu 56: 269-83. [Hebrew]

. 2001. The Babylonian Masora of the Pentateuch: Its Princi-
ples and Methods. The Academy of the Hebrew Language
Sources and Studies 6. Jerusalem: The Hebrew University

Magnes Press. [Hebrew]

. 2011. ‘An Old Manuscript with Babylonian Vocalization
of the Hagiographa’. In Israel: Linguistic Studies in the
Memory of Israel Yeivin, edited by Rafael I. (Singer) Zer and
Yosef Ofer, 129-54. Publications of the Hebrew University
Bible Project 6. Jerusalem: Hebrew University Magnes
Press. [Hebrew]

. 2019. The Masora on Scripture and Its Methods. Fontes et

Subsidia ad Bibliam Pertinentes 7. Berlin: de Gruyter.

The Masoretic Notes in RNL EVR II B 80+ 73

Penkower, Jordan S. 2020. ‘The Biblical Variants between Sura
and Nehardea: Text, Vocalization, Open and Closed Sec-
tions’. Jewish Studies Internet Journal 18. [Hebrew]

Strack, Hermann L. 1897. ‘Uber Verloren Gegangene Handschrif-
ten des Alten Testaments’. In Semitic Studies in Memory of
Rey. Dr. Alexander Kohut, edited by George Alexander Ko-
hut, 560-72. Berlin: S. Calvary & Co.

Weil, Gerard E. 1963. ‘La Massorah Magna Babylonienne des
Prophetes’. Textus 3/1: 163-70.

Yeivin, Israel. 1966. ‘Two Terms of the Babylonian Masora to the
Bible’. Leshonenu 30/1: 25-28. [Hebrew]

. 1968. The Aleppo Codex of the Bible: A Study of its Vocali-

zation and Accentuation. Publications of the Hebrew Univer-

sity Bible Project 3. Jerusalem: Magnes Press, The Hebrew
University. [Hebrew]

. 1973. ‘More on the Traditional Babylonian Term y7’.
Leshonenu 37: 154-56. [Hebrew]

. 1985. The Hebrew Language Tradition as Reflected in the

Babylonian Vocalization. 2 vols. Jerusalem: The Academy of
the Hebrew Language. [Hebrew]
. 2003. The Biblical Masorah. Studies in Language 3. Jeru-

salem: The Academy of the Hebrew Language. [Hebrew]

THE MARGINAL NUN/ZAYIN: MEANING,
PURPOSE, LOCALISATION!’

Vincent D. Beiler

In some early masoretic Bible codices, a large letter resembling
nun or zayin may occur in the margin, often in conjunction with
the marking of gere/ketiv.? The Aleppo Codex does not have this
marking even once, while another illustrious codex, the Cairo Co-
dex of the Prophets (C), has the marking more than 500 times on
about as many pages (Martin-Contreras 2015, 81). This large let-
ter is generally absent or infrequent in codices long cited by
scholars (with the already noted exception of C).* For example,
there are only 76 such letters in the Leningrad Codex (Martin-
Contreras 2015, 88-90) and 42 in Heb.24°5702 (formerly known
as Sassoon 507) (Himbaza 2000, 175). To the best of my

knowledge, no such markings occur in either British Library Or.

' Special thanks to Joseph Habib, for his critical comments and encour-
agement, and to Elvira Martin-Contreras, for her willingness to interact
with and critique my ideas—the paper is better for it.

? For purposes of convenience, I group together all possible types of gere
marking (i.e., gere, ketiv we-la gere, qere we-la ketiv, and ketiv), and shall
refer to them to simply as gere/ketiv. See Yeivin (1980, 56-59).

> Cf. Breuer (1976, 14) for commonly cited codices.

76 Beiler

4445 or JUD 002 (formerly Sassoon 1053).* It has been suggested
that the marking is ancient, perhaps predating the remainder of
the masora magna and parva (e.g., Yeivin 1980, 52). Its distribu-
tion appears to be widespread, being found in Tiberian, Babylo-
nian, and Palestinian manuscripts—this being the primary reason
that the mark is thought to predate the remaining masora (Ofer
2019, 89-91). Yeivin (1980, 52) notes that the letter generally
fell out of use after the 12th century, although Penkower finds
limited instances of the letter in later codices and scrolls (2019).

Scholars of the past and present have offered their opinions
regarding both this signifier and what it might signify. If nun,
perhaps the letter stands for ,») ‘what is read’ (Kahle*) or xno:
‘variant’ (BH?, 51). If zayin, perhaps the letter stands for xn’or
‘uncertain’ (Yeivin 1980, 52). The letter could even be a simple
marking and not a letter at all (again Yeivin). After examining
seventeen diverse codices containing the letter, Himbaza argues
that the sign was probably a nun (2000, 173).

As this sign occurs most frequently in conjunction with
qere/ketiv notes, the letter may signal the existence of an alter-

nate tradition to the reader. The letter does not accompany gere

4 The full name of what was formerly known as Sassoon 1053 is Geneva,
Jacqui E. Safra, JUD 002. Special thanks to Nehemia Gordon, who
kindly provided me with the colour images, and to Jolanda van Nijen
of the Jacqui E. Safra Judaica Collection, who has been instrumental in
permitting scholars access to said images.

° As quoted by Yeivin (1980, 52); Himbaza (2000); Martin-Contreras
(2015). None of these scholars, however, indicate where Kahle is pur-
ported to have said this—nor have I yet succeeded in finding it.

The Marginal Nun/Zayin 77

notes in all instances, however, nor do gere markings—either ex-
plicitly or implicitly—always appear to accompany it (e.g., EVR
II B 1233, 2 Kgs 19.13). In some early scrolls and Ashkenazi and
Italian codices—but not in Oriental codices—the marker denotes
section divisions on which there is disagreement (Penkower,
2019).

Himbaza suggests that the marginal letter serves to alert
the reader to a textual problem (2000, 174). Martin-Contreras
argues that (in the Cairo Codex) the letter is a warning marker,
alerting the reader to an issue in the consonantal text without
explaining it (2015, 88). Penkower (2019) notes that the mar-
ginal letter is employed to mark points of dispute.

A related, but distinctly different sign, nami 31 nun menu-
zeret ‘isolated nun’, also known as 12157 711 nun hafukha ‘inverted
nun’, appears in Bible codices at Num. 10.35-36 (2x) and Ps.
107.23-28, 40 (7x).° While the meaning of this ‘isolated nun’ is
debated, it may indicate that a portion of text is out of place (b.
Shabbat 115b-116a). Lieberman (1962, 38-46) finds a parallel
with certain Greek texts that use an antisigma, i.e., reverse sigma,
to indicate misplaced text (cf. Yeivin 1980, 46-47). Tov (2001,
54-55) believes that the use of the sigma and antisigma pair (what
became our modern-day parentheses) can be seen in
11QpaleoLev’, indicating the long-standing use of the notation in
Hebrew texts to mark wrongly placed verses. It is possible that

our marginal letter is a later outgrowth of the ‘isolated nun’. How-

° Not all codices are in agreement regarding the exact verses in Psalm
107 where the ‘inverted nun’ should appear.

78 Beiler

ever, as the two signs are never confused with one another—de-
spite a potential overlap of meaning—there is no evidence to sup-
port such an assertion. Perhaps the signs are only accidentally
similar.

I shall propose that the purpose of the letter, possibly a za-
yin, was (or became) practical: a means of avoiding certain types
of copyist mistakes when recording gere/ketiv notes. Because the
sign occurs in certain script types more than others, I will also
argue that the notation can function as something of a regional
identifier, although I leave that region to be identified by others.
The explanations offered here are generally compatible with, but

independent from, the explanations cited above.

1.0. Description of Corpus

The data for the present paper are drawn from a database of ca.
15,000 Masorah parva (Mp) notes, taken from 38 early (the 12th
century and prior) codices containing the Former Prophets. The
study is larger than 38 isolated classmarks, however. Apart from
the original 38, there are as many as 43 additional classmarks

containing leaves from one of the original 38.’ To the best of my

” Some of the joins are obvious: the reference ranges, the number of
lines, and the script similarities prevent other conclusions. Other joins
are less certain. The complete list of the 38 classmarks, along with the
potential joins for each, are as follows (a plus sign following a classmark
indicates the presence of possible joins; the join suggestions are listed
in parentheses following the listing of the main classmark; ‘I’ = First
Firkovich collection; ‘I’ = 2nd Firkovich collection): the Aleppo Codex,
the Cairo Codex, the Leningrad Codex, JUD 002, I Bibl. 13/80, I Bibl.
68, II B 24+ (II B 1184, II B 1323, II B 1335, note that several folios of

The Marginal Nun/Zayin 79

knowledge, the study includes all early manuscripts for which I
was able to access images.®

I examined the big four codices as pertains to the Former
Prophets (Aleppo, Leningrad, Cairo, and JUD 002), plus two
lesser-known codices of the First Firkovitch collection, I Bibl.
13/80 and I Bibl. 68 of the library at St Petersburg. The remain-
der are all St Petersburg II B classmarks (that is, from the Second
Firkovitch collection). Every codex from the I Bibl. or the II B
collection was examined, provided it contained at least one of the
predetermined reference ranges (listed below), was sufficiently
‘early’ in appearance (e.g., left justification method did not in-
clude significant letter elongation; see Beit-Arié 2021, 472, nn.
30, 31), and had three columns.° All codices fall under the rubric

II B 24 do not belong), II B 25+(II B 145, parts of II B 210, II B 223, II
B 1197), II B 35, II B 39+(II B 217), I B 43, 11 B 50+(II B 1298, IIB
1349, II B 1379), II B 55, II B 56+(iI B 211, parts of II B 81, most of II
B 71, II B 216), II B 63, II B 70+(II B 212), II B 71(two folios belong
with II B 56+), II B 77 (one section of II B 210, parts of II B 1328, IIB
1345), II B 86+(II B 1405, II B 1406), II B 90, II B 94, II B 99+ (IIB
219, IT B 1269, II B 1325, II B 1326, II B 1339, II B 224, II B 1278), IIB
124, II B 206, II B 927, II B 1160+(II B 1159, II B 1157, I B 1162, IIB
1248, II B 1280, II B 1286), II B 1166+ (II B 207, II B 1247), II B 1167,
IIB 1169, II B 1180+ (II B 1211, II B 1235), II B 1233, I B 1243+(IIB
1255), II B 1270, II B 1272+ (II B 1328), II B 1275, II B 1281+(I B
1337), II B 1285+ (II B 1474), II B 1378+ (one section of II B 24, last
half of IIB 81, II B 134, II B 1336).

8 For example, Gottheil 27 (Breuer’s Codex Lm of the Former Prophets),
of the pen of Samuel ben Jacob, has not been digitised.

° There is one exception. EVR II B 124 is two-column, but exceptionally
early, the colophon (if believed) dating to 946 cE. Regarding the forging

80 Beiler

of Oriental,'° broadly speaking, and have Tiberian vocalisation.
Despite these unifying characteristics, the scripts of the manu-
scripts show considerable variety.

The collation was limited to masora parva notes that fall
within four reference ranges (Jdg. 3-6, 1 Sam. 16-19, 1 Kgs 8-
10, and 2 Kgs 17-20).'’ Masora parva notes falling outside of
these ranges were not considered. For this reason, there are
doubtless manuscripts with instances of the marginal letter that
I did not record. In such manuscripts, however, occurrences of
the marginal letter are demonstrably infrequent. The conclusions
drawn in this paper, therefore, fit best with codices containing

frequent occurrences of the marginal zayin/nun.

1.1. Manuscripts that Use the Marginal Letter

The marginal letter was observed in fifteen of the 38 codices ex-
amined. In only thirteen of the fifteen codices does the marginal
letter occur with high frequency (i.e., the letter occurs adjacent
to an explicitly marked qgere/ketiv note in the majority of q/k in-

stances); we will focus on these thirteen.’ For the purpose of

hand of Firkovich in the colophon of this codex, see Beit-Arié (2020,
202-3). According to Beit-Arié, the actual date is somewhere between
946 and 1036.

" The data are taken from my larger PhD thesis project (2022), where
manuscripts are compared according to their masora parva note collo-
cations.

” These thirteen high-frequency-inclusion codices include: II B 24+, II
B 35, IT B 43, 11B 50+, 1B71+, I B 927, 11B 1166+, II B 1167, IIB

The Marginal Nun/Zayin 81

analysis, I have arranged the thirteen manuscripts into three sub-
types, based primarily upon whole-page comparisons of the man-
uscripts.

In the first six manuscripts, labelled Script A, the folios are
generally crowded, leaving only small inter-column ‘margin’
spaces between the words and lines. The width of the strokes is
proportionally wide when compared against letter height and
width. Despite the wide stroke marks of the calamus, the script
still manages to be slightly calligraphic, as can be observed par-
ticularly in the first three examples (calligraphic: i.e., the pres-
ence of serifs and the use of a calamus with an angled tip).

There is some similarity of Script A with what Olszowy-
Schlanger (2014, 279-99; 2015, 14-20) has labelled ‘South-
Western Oriental’, i.e., Egypt especially, but the similarity is only
partial.'* There is likewise some similarity of these manuscripts
with what Engel (2013, 486-87) refers to as ‘proto-square script’,

1233, IIB 1243+, IIB 1270, II B 1285+, and the Cairo Codex. The two
remaining codices where the marginal letter occurs with lower fre-
quency are II B 63 and IIB 1160+.

'S The lack of congruence does not, however, rule out the possibility
that ‘Script A’ is Egyptian. It merely shows that the script is not wholly
similar to those identified by Olszowy-Schlanger as SW Oriental (Ol-
szowy-Schlanger, personal communication, December 2021). It should
also be noted that the MSS examined by Olszowy-Schlanger are less
formal in appearance than those I examined (in my case, three-column
Bible manuscripts containing full masora parva and magna). To my
mind, this limits the value of a full comparison of the scripts. Similarly,
the corpus of MSS identified as SW Oriental by Olszowy-Schlanger is
small, again limiting full comparisons.

82 Beiler

which Engel describes as “(d)ense texture composed of small let-
ters and small spaces between words and lines.... There are ran-
dom ornamental characteristics, such as stylised tags on the hor-
izontal lines and a decorative curl of the verticals.” As with the
‘South-Western Oriental’ descriptor, however, the similarity of
‘proto-square script’ and the present Script A is only partial. For
these reasons, I will not attempt to pinpoint the likely point of
origin for these MSS (but see below, 85.3).'*

Figure 1: Script A
EVR II B 1270, p. 15

wore. ‘eet SB SPH 4 pranit
sei yeasty |. on

i e a

—_ ; es Pon) rer | nl Peek ores aa
EVR IIB 35, p. 103

Torre wrayer us : —
serie zt aay Dray TP ie
awa sites. || THT

EVR IIB 1243+, p. 10

Bis ie THM wap wes
came} | BS TINAS INA how
oa =F pm .- - - a: Be

™ Tt should also be noted that my descriptions are necessarily cursory—
and thus provisional; palaeography is complex, and each script type, as
I have identified them (especially Scripts A, B, and D) deserves a much
longer treatment than I am able to provide in this article.

The Marginal Nun/Zayin 83

EVR IIB 1285 + (here, of II B 1474), p. 5

asain 15 igi $5 “ws egene
NDINAINTWHOB VS RAM PDN SII
7 77 * it aS a ! e.

EVR IIB 927, p. 51

ura fT! Ronmassip ies TR
DOWD ~ WY DYTNSIN SOY TDN WONT

EVR II B 1233, p. 20

Script B,'° containing four examples, shows a similarly wide
letter stroke in keeping with that of Script A, but, rather than
appearing wide throughout, the horizontal lines tend to be wide,
and the vertical lines narrow. The letters in Script B are larger
than in Script A, and are of the sort that one encounters in espe-

cially “heroic” productions (the first three examples especially).

'S As the present data set yields only four codices for Script B, it may be
helpful to note other codices of this script type, so that the alert reader
may compare the various codices. These include the Washington Pen-
tateuch, II B 19, II B 20, and II B 1021—all contain the marginal letter
and have at least 60 pages (30 leaves) preserved. Classmarks of Script
B that are too short to find gere/ketiv type notes include II B 1064, IIB
1065, II B 1067, II B 1070, II B 1296. Extensively preserved codices
appearing to belong to Script B, but that I have not had opportunity to
examine beyond a single leaf, include Gottheil 6/Ms. FR 9-005 (note
that Ms. FR 9-005 combines Gottheil 5 with Gottheil 6) and Gottheil
18/C3.

84 Beiler

In many respects the script of these four examples is similar to
Script D (below, in due course), with the exception of the wide
horizontal strokes.

When comparing paratextual features, however, the simi-
larities between Scripts A and B are especially marked (e.g., se-
darim markers are similarly formed; the masora parva gimel is of-
ten triple-dotted’®), suggesting that the scripts may have similar-
ities that extend beyond the presence of the marginal letter’” and
the wide letter strokes. The marginal nun/zayin in these MSS is
very similar in appearance to that in II B 1270, II B 1233, and II
B 1243+ of Script A, again suggesting some overlap in their cen-

tre(s) of production.

16 The supra triple-dotting of the masora parva gimel, on which nothing,
to my knowledge, has been written, appears to be an alternate form for
marking 3x. In some MSS, only the masora parva gimel is triple-dotted.
In other MSS, the triple-dotting feature sees wider distribution, partic-
ularly with letters having a flat roof, such as heh or bet, or in two-digit
numerals (e.g., 7° or 73). The feature is not necessarily ubiquitous within
a given MS, but the triple-dotted gimel and the single-dotted gimel gen-
erally appear to have been formed by the same hand. Sometimes gimels
of identical shape appear on the same page, one triple-dotted and one
single-dotted. Double-dotting over single letters, a much rarer feature,
is generally equally distributed over all letters (cf. BL Or. 9880; Oxford
MS heb. b.17/1).

” Cf. Penkower’s (2021, 160-61) article comparing codicological and
palaeographical similarities between the Washington Pentateuch and
the Cairo Codex.

The Marginal Nun/Zayin 85

Figure 2: Script B
Cairo Codex, p. 273

vara wns
et shania wsindin eae?

ap
n37

EVR IIB 24+, p. 130

movie ADOY pe

EVR IIB 1166+, p.8
em ToMMON To: «— oaay
Pum aanmamamen St oops

There remain three manuscripts where the marginal letter
occurs frequently. I see little in these scripts from a paratextual
or script standpoint that link them to Script A or Script B in any
meaningful way—or to one another. Nonetheless, the marginal
letter appears in these manuscripts with regularity, and I record
that fact here.

86 Beiler

Figure 3: Script C

EVR II B 1167, p. 6
tsp i= * ieee Ld “<7 "yo rE" ~

Wise: ania teen Raat
moi > BRE TY

EVR IIB 43, P. 24

> el Sd

Si DP yy I OV ig 185
Api Fens

wire

EVR IIB 71+, p. 103

1.2. Manuscripts that Do Not Use the Marginal Letter

Our large, marginal letter was not found in 23 of the 38 manu-
scripts (Figures 4 and 5).’® Although the manuscripts lacking the
marginal nun/zayin show script variations, certain trends are

readily observed. Most notably, fourteen codices contain a script

18 NB, some of these 23 MSS still may contain infrequent occurrences of
the marginal letter (e.g., the Leningrad Codex) not found within the
stated reference ranges (Jdg. 3-6, 1 Sam. 16-19, 1 Kgs 8-10, and 2 Kgs
17-20). One should distinguish between studies where the letter occurs
or does not occur with high frequency (the salient point of difference in
the current paper) and MSS where the letter does or does not occur full
stop (cf. Himbaza 2000).

The Marginal Nun/Zayin 87

style found in some of the earliest dated Bible codices in our pos-
session, e.g., EVR II B 17 (930 CE), EVR II B 10 (946 cE), and EVR
II B 39 (989 ce).'? Many scholars will recognise this script due its
congruity with the Aleppo Codex. The fourteen manuscripts are
included below as Script D.

Figure 4: Script D

Aleppo Codex

EVR IIB 39+, p. 10

'® Dated examples of this script type, some already mentioned above,
may be found in Beit-Arié’s Specimens of Mediaeval Hebrew Scripts, Vol-
ume I: Oriental and Yemenite Scripts (1988: esp. plates 4, 5, 10, 11, 12,
15, 16, 19).

88 Beiler

EVR IIB 99+, p. 34

EVR IIB 70+, p. 6

NaN Aum:
ea “un
Arias MPT DATS 73)

EVR IIB 77+, p. 20

EVR II B 1272+, p. 6

n° Oenias 4 Gyan
nD TRANS ey

The Marginal Nun/Zayin 89

EVR II B 1378+, p. 19

EVR IIB 1169, p. 6

my tepeprneensin Awan
x 2 “LADY PD. oie ate

EVRIIB 55+, p.7

EVR IIB 56+, p. 14

= Dent % 7 seep TN nest

EVR Bibl. I 19a/Leningrad Codex

te
SDA a
01 a
wx «86 SHR TDD > a

90 Beiler

The final three of the above manuscripts do not fit into the
Script D category as neatly as the others, but their similarities to the
foregoing eleven are nonetheless remarkable. In sum, the above
fourteen manuscripts show considerable congruence, presenting a
script subtype that has not appeared in manuscripts with frequent
attestations of the large, marginal letter (Scripts A-C).

There remain nine manuscripts (Figure 5, below), which do
not employ the large, marginal letter. These MSS are of several
types. One classmark, EVR II B 90, shows Sephardi influence mixed
with some similarity to Script D. Several manuscripts have very
small and fine writing (e.g., I Bibl. 68, II B 206, II B 94, JUD 002),
making script comparison less productive. EVR II B 1275 is similar
to Script A. Clearly, neither script categories nor scribal practices

were entirely fixed. These final nine manuscripts are listed below.

Figure 5: Final nine codices without the marginal letter

IL B 124, p.17

mn TOR 1 3X A | ISISINN1:
Vi INT:

benrawhmam apts

11B 1281+, p.8

nrako sonwha meg Li

apy aay ann) “yon

~ ro

The Marginal Nun/Zayin 91

I Bibl. 68, p. 5

Sy: mpd “DAE NTIS oma iy —y
re sS370 MAb PATIN ena 23 - _
ine om aye» i + te Sane eee 7 :

JUD 002 (formerly Sassoon 1053)

= ate p. 6
4 cian rer jenn

: s
famis® A Leacer

II B 90, p. 6

hig = hyMUEAWIND mY
7 ~—_ wy

11B 1180+, p.9

92 Beiler

Although the above descriptions provide only the briefest
of an overview, we are left with some unmistakable patterns.
First, codices of the script style of the Aleppo Codex et al. (Script
D) are unlikely to have the marginal letter with any frequency.
Secondly, in the majority of instances, the scripts with the mar-
ginal letter contain paratextual features not found in Script D.
These data support, for example, the argument that the Cairo Co-
dex was not from a ben Asher centre of production (cf. Penkower
1990; Beit-Arié et al. 1997, 28-29).

2.0. Is the Large, Marginal Letter a Final Nun?

In a visual inspection of the present codices, it is difficult to de-
cide if the marginal letter is a zayin or a final nun. In the main
text, context, rather than letter shape, is frequently determina-
tive. As there are no other large letters in the margin against
which our letter may be assessed—and because the large letter is
not part of a word—we must rely upon minor differences of
sometimes questionable significance. On the basis of the codices
that Himbaza (2000, 174) examined, he concluded that the mar-
ginal letter was a nun. Likewise, in the manuscripts surveyed by
Penkower (2019) a nun seemed likely. However, the present evi-
dence is only in partial agreement with those assessments. To my
eye, in only nine of the fifteen manuscripts does the marginal
letter appear more likely to be nun than zayin. My readings are

open to debate, admittedly, and we lack the space to do extensive

The Marginal Nun/Zayin

comparative work in each manuscript. Still, the letter is not nec-

essarily a nun—or, perhaps, was not considered to be a nun at all

times or in all regions.

Table 1: Final nun and zayin comparison

Marginal ; : Most
ID Final nun zayin
letter resembles
, .
II B 63 VY) WT \
‘ ' 9 j nun
(p. 81) ’ ' \ | AA
4 rs
1B 50+ ‘ L
ah a
(pp. 146, 7 tal é y " Y zayin?
149) a ‘ *
/
IIB 24 7 HATH: a
(pp. 103, zayin?
4 —
110) Wy a ie
's Sen Nun h
t i
II B 1167 . .
(56) MINS Py zayin
a ee
\ - .
€ zayin (nun in
IIB 35 i main text is
(p. 92) D’ 1990 markedly
? longer)
;

Il B 43
(p. 5)

II B 1270
(p. 5)

II B 1233
(p. 5)

I. B 1243+
(pp. 5, 9)

II B 927
(p. 35)

II B 1160+
(here, of II
B 1248)
(p. 11)

Il B 1166+
(p. 17)

Beiler

; 5
a MM
i §
.
vo Te
ay TS MT
j ‘ ied

ns AMX

nunt fzayin

nun?

zayin?

nun

nun?

nun

The Marginal Nun/Zayin 95

II B 1285+
(1474)
(p. 5)
nun
IWB71+
(letter is obvi-
(p. 91)

ously long)

y Ia

Cairo Codex $ FS 7 " ; | ;
7 zayin

(p. 273) ri f ose ¥ ly’

Left unaddressed in the nun or zayin discussion is a rather
troublesome question. Namely, if the letter is a nun, why should
the final form be used to mark an abbreviation? Why not simply
use a non-final nun as one would do to indicate the number fifty?
As noted above, Kahle suggested that the final nun stood for ;»"7),
a solution which Yeivin dismissed as “astonishing” (1980, 52). I
would tend to agree. To the best of my knowledge, there is no
precedent for the use of a final form to indicate a non-final letter
within the masora parva. For example, the abbreviation for pp105
‘verses’ is 5, never 4; the abbreviation for 1m ‘from’ is A, never 0.”°

The exception to this rule occurs only in mnemonics. In
Deut. 30.16, the proper sequence for the three-word phrase rniya
roawm rnpmi ‘commandments, and statutes, and judgements’ is in-

dicated by BHS as 4px. In manuscripts, however, the pe is always

2° Instances of A indicating ‘from’ can be found in virtually every codex.
Instances of 51 indicating ‘in the verses’ are less frequent—but see, for
example, the Cairo Codex: Jdg. 5.13 (2x), 23, 30 (2x).

96 Beiler

written in its final form, namely 4p¥.2! Similarly, the mnemonic
ji appears at Ps. 25.7 in the Leningrad Codex. Here, & = 738 or
m>x, f = nxt; j = &3. Once again, a non-final letter is written as a
final letter. In both cases, the reason the final letter of the three-
word mnemonic takes the final form is due to its position within
the mnemonic. As a stand-alone abbreviation, however, there is
no reason for a non-final letter to be written as a final letter.

The most reasonable path in identifying the marginal letter,
then, is to assume that it cannot be a nun and is therefore either
a zayin or simply a marker of unknown meaning which happens
to resemble a final nun. With this in mind, we turn to an angle of

the problem that has not received treatment in the literature.

3.0. Masora parva Notation as a Two-Step Process

Although our marginal letter is often the same size as the main
text, there are a minority of instances where the letter is more
nearly the smaller size of the surrounding masora parva. In some
codices (e.g., EVR II B 1167), the marginal letter is characteristi-
cally small. In others, the size varies from instance to instance
(e.g., EVR II B 43, EVR II B 973, EVR II B 1243+, EVR II B
1285+ ). These codices with variable-sized marginal letters pro-
vide us with an important bit of insight regarding the order of
operations in the writing of a codex. Namely, when our letter is

large, its central position in the margin indicates it was added at

?1 In my database of Torah MSS at Deut. 30.16, a final pe is written 9x:
Washington Pentateuch (later additions), JUD 002, II B 10, Vat.evr.448,
II B 59, II B 96, II B 74, II B 18, and II B 158. No occurrences of a non-
final pe were observed.

The Marginal Nun/Zayin 97

a point prior to the gere note and/or the surrounding masora
parva notes. When the marginal letter is more nearly the size of
the remaining masora parva, however, its position indicates that
it was added at the same time as the surrounding masora parva
notes. These statements can be proven through a careful exami-

nation of the following images.

3.1. Letter Size as Evidence of a Two-Step Process

Figure 6: II B 927, p. 41

sd 2 riled * Law

5 a oer 4 ne zsh é 133 ; e/ “i 4 (

A PAL 22 ee 2'o® awe fa. Rae SS Geri

1
|
i,

In the above image are three masora parva notes. They concern,
in verse order, the words np? ‘is taken’, rywi/>wn ‘their rulers’,
and 35°77 ‘make them howl’ (Isa. 52.5). In situations such as this,
where multiple masora parva notes occur on the same line, the
masora parva comments generally are organised according to
verse order. This means that the masora parva note for n7? should
occur farthest to the right, the masora parva note for 35°77 should
occur farthest to the left, and the masora parva note for rywh/wn
should occur somewhere between the above two.

In the present instance, however, there are some problems.
Our large letter occurs farthest to the right, and its associated
qere comment () *), is squeezed alongside, below and to the left.
The gimel associated with np), rather than occurring farthest
right, now occupies the second spot. The explanation for this re-

versal in the order of masora parva notes is not hard to find: the

98 Beiler

large letter was written prior to and without regard for the re-
maining notes. Rather than erase the large letter and start over,
the scribe of the smaller masora parva notes elected to place the
gimel in the second place instead of the first. An alert reader
would perhaps have had no trouble sorting out these comments.
The fact remains that this reverse ordering of the notes is highly
atypical in the present corpus.

Contrast the above example with the following one, also
from EVR II B 927, where the marginal letter is the size of the
remaining masora parva comment. Here, one can see how the en-
tire comment is integrated, our marginal letter occurring slightly
to right of centre in the margin space, allowing for the comforta-

ble addition of the remaining portion of the note.

Figure 7: II B 927, p. 10
ad = ed 4 VAY | ry os chon VAY ; : ie

Saeryameienot DF

This pattern holds for all codices examined above. Where the mar-

ginal letter is large, it tends to occupy pride of place, in the centre of
the margin space or slightly to right of centre—perhaps in anticipa-
tion of an eventual gere/ketiv comment. When the letter is smaller, it

is more integrated, space-wise, with surrounding masora parva notes.

3.2. Letter Placement as Evidence of a Two-Step Process

In codices where the marginal letter is large, it almost always
occurs level with the word(s) being commented upon. That is,

the letter is hung from the line in the same way as the main text.

The Marginal Nun/Zayin 99

Figure 8: II B 24, p. 130

When the marginal letter is smaller, however, it may appear
above the line of the main text being commented upon in the
following manner:

Figure 9: II B 43, p. 24
‘ \:'2Ua rare : a tal

+
= 1 OD
Fe satsioe @ PURO ? amma _—_

Although the defective spelling of wx ‘man’ (2 Sam. 14.19)

is being commented upon in the above image, our small marginal
letter sits just above the headline of the referent, rather than
hanging just below it. Why? It appears that the scribe, consider-
ing the note as a whole, chose to align the centre of the note most
nearly with the proper line of the main text. Contrast the above
example from Figure 9 with the former example in Figure 8,
where the large marginal letter was written level with the main
text with no regard for eventual placement of smaller masora
parva notes. Once again, in codices that contain the large, mar-
ginal letter, the writing of the masora parva appears to have been

a process consisting of at least two steps.”

22 An exception may be found, for example, in the Cairo Codex at 1 Sam.
17.23. In this instance, the gere note p ninnynn was written first and our

100 Beiler

3.3. Ink Differences as Evidence of a Two-Step

Process2*

We can reach a similar conclusion when comparing the ink of the

large, marginal letter with the remaining masora parva note.

Figure 10: Washington Pentateuch, Exod. 22.26
ea dca
noe Senate
yp

Bae @ fae

letter, of small size, is squeezed in beside the gere comment, almost as
an afterthought. This example, nonetheless, proves the general rule: a
large nun/zayin is written prior while a small nun/zayin is written later.

23 NB, images of the Washington Pentateuch and Schgyen 1630 were
chosen due to the availability of colour images; they are not part of the
current data set, which are comprised of black and white images with
only three exceptions (Aleppo, Jud 002, Leningrad, the first two of
which contain no instances of the marginal letter). Also note that 75 of
the 76 occurrences of the marginal letter in Leningrad occur in the Writ-
ings, the sole exception appearing in Jdg. 20.13 (Martin-Contreras
2015, 88-89). The lopsided distribution of the marginal letter in Lenin-
grad remains unexplained, although the presence of a secondary scribe
(or scribes) seems likely (cf. Himbaza 2017, 355-68).

The Marginal Nun/Zayin 101

Figure 11: Leningrad Codex, Job 26.12

iit 60 ot ae
' aye
Ss Moors :
*
A da be

In the Washington Pentateuch, the ink of the large letter is in
relatively good condition, matching the ink of the main text. The
qere comment, by contrast, appears to have flaked off. In the case
of the Leningrad Codex, the reverse has occurred. The main text
and the marginal letter are dim (the main text obviously has been
reinked) while the ink of the smaller masora parva hand remains
intact.

Something similar has happened with Schgyen 1630. Alt-
hough the image quality is poor, one can still see a black ink used
for main text and the large, marginal letters. A (now) reddish ink
was used for the remaining Mp notes, including those accompa-
nying the marginal letter.

Figure 12: Schgyen 1630, Zech. 14

—
on ow

“4

tal
:
AG

aaa
Y4echiyds

st +:

. ee ee —_ ane
In sum, it appears that the marking of the masora parva

notes was a process of at least two stages. First, probably at the

102 Beiler

time of writing of the main text, a large letter was recorded in
the margin at qere/ketiv-type instances. Secondly, at a time after
the large letter had been inserted, the remaining masora parva
was written. In instances where the large marginal letter was in-
itially ‘missed’, the later scribe sometimes still elected to write it
in the margin, albeit now with the hand size of the remaining

masora parva notes.

4.0. Does the Large, Marginal Letter Always

Indicate qere/ketiv?

It is the view of Martin-Contreras that in C the marginal letter
does not necessarily indicate gere/ketiv. Supporting this claim are
90 instances in C where the marginal letter appears without a
qgere/ketiv type note alongside (2015, 83-89).”* However, the as-

sertion of Martin-Contreras regarding the putative independence

4 My calculations are based upon the data of Martin-Contreras (2015,
83-87). To arrive at 90 occurrences, I omitted instances said to be
marked for ketiv or gere or some combination of the two. This yielded
93 occurrences. Of these 93, three appear to have been entered in error
(marginal letter not visible: Jdg. 21.20; gere note present but marked as
not present: 2 Sam. 22.23; 2 Kgs 11.1), bringing the number of instances
to 90. The references are as follows. From §1, 2x: 1 Sam. 9.26; 2 Sam.
14.11. From §2.1, 12x: 1 Sam. 24.19; 25.3, 8; 2 Sam. 3.3; 1 Kgs 21.21;
2 Kgs 13.6; Jer. 19.15; 32.35; 39.16; Ezek. 16.25; 23.43; Mic 1.15. From
§2.2, 75x: Josh. 6.7; 7.21; Jdg. 9.8, 12; 13.17; 17.2; 1 Sam. 15.16; 24.5;
25.34; 28.8; 2 Sam. 1.16; 10.9, 17; 11.1, 24; 12.1, 13.8; 14.11; 21.9;
23.20; 1 Kgs 1.27; 2.24; 6.16; 8.26; 9.25; 12.3, 21; 18.36; 22.13; 2 Kgs
7.12; 8.21; 11.15; 14.2; Isa. 23.12; 26.20; 42.24; Jer. 1.5; 2.33; 3.4;
4.19; 5.7; 10.17; 15.16; 22.23; 31.21, 39; 34.11; 41.17; 42.20; 46.11;
48.44; 49.28; 51.13; Ezek. 4.6; 9.5, 8; 16.13, 18, 22, 31, 31, 43, 47, 51;

The Marginal Nun/Zayin 103

of the marginal letter appears less certain when comparing C to
other early codices. Of the 90 instances where C contains a mar-
ginal letter but does not mark qere/ketiv (see above, n. 24), 79 of
them are explicitly marked as qgere in the Leningrad Codex.”° Of
the remaining eleven, I was able to find explicit qere marking in
eight instances,** leaving only three examples?” where a gere
marking was not found. In other words, even if a note is not
qere/ketiv in C, this is not to say that the note was not considered
to be so by masoretic scribes more generally.

Moreover, 75 of the above 90 instances in C where the mar-
ginal letter occurs without explicit marking of gere occur in con-
junction with wn ‘superfluous’, a term regarded by Yeivin as oc-
curring in “qere/ketiv situations” (1980, 94).

Penkower (2019) provides numerous examples, primarily
in Torah scrolls, where the marginal letter was not intended to

preserve gere/ketiv, but instead marked section division disagree-

18.28; 23.14, 42; 27.3; 29.4; 36.16; Hos. 4.6; 10.14; Mic. 1.3, 10, Zech.
1.4. From §§2.3-2.6, Ox. From §2.7, 1x: 2 Kgs 5.25.

5 The eleven instances not marked as gere in either L or C are: 1 Sam.
25.3, 8; 2 Sam. 10.17; 11.1; 12.1; 1 Kgs 9.25; 2 Kgs 8.21; Ezek. 9.8;
Hos. 4.6; 10.14; Mic. 1.15.

76 1 Sam. 28.8 (I Bibl. 68); 2 Sam. 11.1 (II B 43); 2 Sam. 12.1 (11 B 1255-
part of II B 1243+); 1 Kgs 9.25 (II B 35); 2 Kgs 8.21 (II B 35); Ezek. 9.8
CII B 24); Hos. 4.6 (II B 50); Hos. 10.14 (II B 50).

?7 No gere note found: 7328 ‘Abigail’ (1 Sam. 25.3); nax2n ‘to Helam’ (2
Sam. 10.17); °ax ‘I will bring’ (Mic. 1.15). There may be codices where
qere is marked at these locations; I have not yet succeeded in finding
them.

104 Beiler

ments. The few codices cited by Penkower, however, are (a) en-
tirely Ashkenazi or Italian and (b) generally later than the present
corpus. As the use of the large marginal letter to indicate disputed
section divisions was not found in the present study focusing on
Oriental codices, it appears that the letter’s meaning could have
varied from region to region (Ashkenazi vs. Oriental) or medium
to medium (Torah scrolls vs. codices). It is also possible that the
signs, although visually similar, were not understood by the
scribes who wrote them as being identical.

In the present corpus, the marginal letter is almost wholly
associated with gere/ketiv. Where the letter appears without the
explicit mention of gere in a given codex, the word has a gere
marking at the same reference in a codex elsewhere within the
corpus.

There are also instances where the gere is implicitly pre-
sent, even when no mention of it is made in the margin. For ex-
ample, in EVR II B 50 at Jdg. 4.11, the marginal letter appears
by itself; there is no qere marking in evidence. The word being
commented upon, D°3ip¥3/o"1yy1 ‘at Zaanaim’, has the ketiv
spelling, but the qgere vocalisation. This indicates that the qere,
although perhaps not explicitly marked, was certainly known to
the scribe who wrote the note. These data point towards a usage
of the large marginal letter in Oriental codices that is wholly fo-

cused on gere/ketiv.

The Marginal Nun/Zayin 105

5.0. Suggestions Regarding the Letter’s Meaning,

Use, and Localisation

5.1. The Marginal Letter as Zayin

As demonstrated above, there is insufficient evidence to decide if
the marginal letter is a symbol or an actual letter. If the marginal
letter began its existence as an actual letter with an actual mean-
ing, however, we should take a fresh look at possibilities that
begin with the letter zayin.*® Yeivin has suggested xnvrv"r ‘uncer-
tain’, based upon a comment found in EVR II B 10:

wean on WPI 7/7 K|T 7 75 WNIT RNP TNA NMDA PK
pay prba npn xoot pPRi ns oN

These are the words in the Torah where t/j is written beside
the column, and a dot is marked above a word or letter.
That word or letter is uncertain, and there are different
opinions on it.”

Penkower notes, however, that the ensuing list of catchwords in
EVR II B 10 pertain to space break disagreements for Torah
scrolls (2019, 145; cf. Dotan 2007, 616). In other words, the com-

ment in EVR II B 10 does not appear relevant when considering

8 There are numerous instances where the letter is undoubtedly final
nun-like (cf. Penkower 2019), but as these manuscripts are generally
later, one should be careful before assuming that the meaning of the
letter remained static while the shape of the letter underwent modifica-
tions.

?° The identity of the symbol in this example is not clear-cut, necessitat-
ing the dual entry of the note as 1/j.

106 Beiler

non-space break type usages such as those found in the present
corpus.

We might also consider other options, such as 131 ‘remem-
ber!’ or pait/xat ‘pair/pairs’, either of which would serve as a re-
minder to the reader that what is on the page is different from
what should be read. This latter suggestion is especially conso-
nant with the masoretic ethos, being part and parcel of the mas-
oretic project as seen most famously in Sefer Okhlah ve-Okhlah. It
is also precisely the kind of note one would expect in a text where

a known difference (ketiv and gere) could be found.

5.2. Why Was the Marginal Letter Large?

Why was the marginal letter large, especially in codices where
the marginal letter occurred the most frequently? Here, I would
suggest two reasons, one historical and the other practical.

The consensus is that the marginal letter is early, that is,
probably predating the remaining masora parva notes (e.g., Ofer
2019, 89-91). For scribes at this stage of transmission (now lost
to us), there was little reason for the letter to be small. The scribe
simply wrote the letter in the margin using the calamus he, or
she, was already using, establishing the pattern that the letter be
large.

As time went on, masora parva comments and gere/ketiv
comments also were added to the margins of Bible codices. This
made the marginal letter somewhat redundant (Penkower 2019).
That is, the marginal letter was merely signalling what the gere

made explicit; it did not serve an additional function. The writing

The Marginal Nun/Zayin 107

of the letter continued to serve a practical purpose, however,
which may have extended its use.

Scribes recording the masora enjoyed freedom to include
whichever notes they preferred (i.e., Yeivin’s Rule). The only ex-
ception was where the traditions for ketiv and gqere differed. In
these situations, both traditions, if known, must necessarily be
recorded. To ensure that these differences were not missed, it was
convenient to write our letter at the time when the consonants of
the main text were written. This would provide practically fail-
safe assurance that the location of the ‘pair’-—or whatever the
sign meant—would receive comment by the later scribe(s) when
the smaller masora parva notes were written.

Of course, instances would have arisen where the large let-
ter was inadvertently missed, or where a later scribe judged that
the large letter should not have been written in the first place.
This is to be expected, as the exact assemblage of qere notes varies
from codex to codex (cf. Ofer 2008; 2019, 92-93). In instances
where the marginal letter was not included during the initial en-
try process, scribes of the remaining masora parva did not neces-
sarily feel compelled to write this letter, its primary usefulness

having already passed.*°

°° There is an additional explanation for the letter’s size which merits
consideration. Namely, since one likely use of a model codex was to
ensure an accurate public reading in the synagogue, the presence of the
marginal letter (particularly if it was of a size that could be readily spot-
ted) would assist the proofer in correcting the reader at the very point
where the reader was most likely to go wrong: the pronunciation of the
qere. Many thanks to Estara Arrant for this suggestion.

108 Beiler

5.3. The Marginal Letter and Localisation

Finally, we arrive at the issue of localisation. There are some tan-
talising clues to suggest centres of scribal activity where the mar-
ginal letter was more likely to occur. In the manuscripts with col-
ophons that Himbaza (2000, 187) examined, an Egyptian origin
seemed the most likely. In the manuscripts I examined, it was
observed that several manuscripts show some similarity in their
features with SW Oriental scripts. Recognizing that uncertainties
remain, I do not insist upon this association, but suggest it as a
line of inquiry that merits further research.

Regardless, it remains abundantly clear that scripts resem-
bling that of the Aleppo Codex and manuscripts like it are the
least likely of all codices to have the large marginal letter with
any frequency. This provides us with, at minimum, a negative
definition. Namely, codices that contain frequent attestations of
the large marginal letter are not from centres of production asso-
ciated with NE Oriental script of the ‘Tiberian’ type (i.e., Script
D, above).

In the present corpus, the only manuscript with the mar-
ginal letter containing a colophon is the Cairo Codex—whose col-
ophon claims that it was written in Tiberias.*' As the Cairo Codex
shows some significant differences in paratextual features—not

to mention vocalisation differences (cf. Yeivin 2003, 13-19) with

31 As stated above, the Leningrad Codex is here excluded, as its sparing
use of the marginal letter does not occur with the reference ranges un-
der examination (for the occurrences in L with references, see Martin-
Contreras 2015; Himbaza 2000).

The Marginal Nun/Zayin 109

other codices known to have originated in Tiberias and Jerusa-
lem—it remains possible that Egypt is more probable for the writ-
ing of C than Palestine. But, as a large number of 10th-12th-cen-
tury Bible manuscripts that have survived to the present origi-
nated in, passed through, or ended up in Egypt, suggesting an
Egyptian origin is not entirely illuminating. As Olszowy-
Schlanger (2015, 14-20) notes, competing intellectual centres
may sometimes be found within the same city (cf. Engel 2013,
488). A more fruitful line of inquiry may be to identify paratex-
tual feature distributions, which, in turn, could suggest scribal
schools. Once the ‘schools’ become better understood, we will be
in a firmer position to posit likely localisation—or to discuss it in
a more meaningful manner.”

Despite these necessary qualifiers, it remains likely that the
writing of the marginal letter, for the time period of the manu-
scripts in question (10th-12th century), was a scribal practice
more associated with Egypt than elsewhere. It may even be that
the letter fell out of use precisely due to the increased prestige of

the Aleppo Codex and codices similar to it.

»? For example, script sub-type, the large marginal letter, the triple-dot-
ting of certain masora parva notes, and the use of the marginal 7
‘well/good/correct’ with no additional qualifiers tend to occur in tan-
dem. Prima facie, this would indicate localisation. For other paratextual
features that may occur in tandem, see Penkower (2021, 160-61).

110 Beiler

6.0. Summary

The present article considers the recurrent use (or absence) of a
particular marginal letter in 38 early Bible codices. In these co-
dices, the use of the letter is limited to certain script subtypes,
suggesting that regional difference and/or scribal school heavily
influenced the letter’s usage, particularly in the codices where
the letter appears with high frequency. The placement and size
of the letter suggest a two-stage method of masora parva note en-
try (first the large letter, and later the smaller writing). Finally,
it may be unhelpful to identify the symbol as a final nun. Instead,
the letter may be better understood as a scribal sign (either zayin
or a sign of unknown meaning) written to indicate an alternate
tradition (qere/ketiv) that the scribe of the main text was at spe-

cial pains to record.

References

Beit-Arie, Malachi. 2020. ‘Supplement: The Forgery of Colophons
and Ownership of Hebrew Codices and Scrolls by Abraham
Firkowicz’. Fakes and Forgeries of Written Artifacts from An-
cient Mesopotamia to Modern China, edited by Cécile
Michel and Michael Friedrich, 195-205. Berlin-Boston: De
Gruyter.

. 2021. Hebrew Codicology: Historical and Comparative Ty-
pology of Medieval Hebrew Codices Based on the Documenta-
tion of the Extant Dated Manuscripts until 1540 Using a Quan-
titative Approach. Jerusalem: The Israel Academy of Sci-

ences and Humanities.

The Marginal Nun/Zayin 111

Beit-Arié, Malachi, Colette Sirat, and Mordecai Glatzer. 1997. Co-
dices hebraicis letteris exarati quo tempore scripti fuerint ex-
hibentes, Vol. I. Turnhout: Brepols.

Beit-Arié, Malachi, Edna Engel, and Ada Yardeni. 1987. Speci-
mens of Mediaeval Hebrew Scripts, Vol. 1: Oriental and
Yemenite Scripts. Jerusalem: The Israel Academy of Sci-
ences. [Hebrew]

Breuer, Mordechai. 1976. The Aleppo Codex and the Accepted Text
of the Bible. Jerusalem: Mosad Harav Kook. [Hebrew]
Dotan, Aron. 2007. ‘Masorah’. Encyclopaedia Judaica, 2nd edi-
tion, edited by Fred Skolnik, et al., XIII:604-56. Jerusalem:

Keter Publishing House.

Engel, Edna. 1998. ‘Style of the Hebrew Script in the Tenth and
Eleventh Centuries in the Light of Dated and Datable Geni-
zah Documents’. Te‘uda 15: 365-410. [Hebrew]

. 2013. ‘Script, History of Development’. In Encyclopedia of

Hebrew Language and Linguistics, edited by Geoffrey Khan,
et al., III:485-502. Leiden: Brill.

Ginsburg, Christian D. 1880. The Massorah Compiled from Manu-
scripts. 3 vols. London: no publisher.

Himbaza, Innocent. 2000. ‘Le Nin Marginal et la Petite Massore’.
Textus 20: 173-91.

. 2017. ‘La diversité des sources du manuscrit de Lenin-
grad B19a’. Semitica 59: 355-68.

Kahle, Paul. 1927. Masoreten des Westens. Stuttgart: Kohlhammer.

Kelly, Page H., Daniel S. Mynatt, and Timothy G. Crawford. 1998.
The Masorah of Biblia Hebraica Stuttgartensia. Grand Rapids,
MI: William B. Eerdmans.

112 Beiler

Lieberman, Saul. 1962. Hellenism in Jewish Palestine: Studies in
the Literary Transmission, Beliefs, and Manners of Palestine in
the I Century B.C.E.-IV Century C.E., 2nd edition. New York:
Jewish Theological Seminary of America.

Martin-Contreras, Elvira. 2015. ‘The Marginal Nun in the Masora
of the Cairo Codex of the Prophets: Use and Function’. Vetus
Testamentum 65: 81-90.

Ofer, Yosef. 2008. ‘Ketiv and Qere: The Phenomenon, Its Nota-
tions, and Its Reflection in Early Rabbinic Literature’. Lesho-
nenu 70: 55-73. [Hebrew]

. 2019. The Masora on Scripture and Its Methods. Fontes et

Subsidia ad Bibliam pertinentes 7. Berlin: De Gruyter.

Olszowy-Schlanger, Judith. 2014. ‘On the Hebrew Script of the
Greek-Hebrew Palimpsests from the Cairo Genizah’. In The
Jewish-Greek Tradition in Antiquity and the Byzantine Empire,
edited by James Aitken and J. Carleton Paget, 279-99.
Cambridge: Cambridge University Press.

. 2015. ‘Manuscrits hébreux et judéo-arabes médiévaux’.
Annuaire de l'Ecole pratique des hautes études (EPHE), Section
des sciences historiques et philologiques 146: 14-20.

Penkower, Jordan. 1990. ‘A Pentateuch Fragment of the Tenth
Century Attributed to Moses Ben-Asher (MS Firkovicz B
188)’. Tarbiz 60: 355-70. [Hebrew]

. 2019. ‘The 12th-13th-Century Torah Scroll in Bologna:

How It Differs from Contemporary Scrolls’. In The Ancient

Sefer Torah of Bologna: Features and History, edited by
Mauro Perani, 135-66. ‘European Genizah’: Text and Stud-
ies 59. Leiden: Brill.

The Marginal Nun/Zayin 113

. 2021. ‘An Eleventh-Century Eastern Masoretic Codex of
the Pentateuch’. Textus 30: 152-70.

Tov, Emmanuel. 2001. Textual Criticism of the Hebrew Bible, 2nd
edition. Minneapolis, MN: Augsburg Fortress-Assen: Kon-
inklijke Van Gorcum.

Yeivin, Israel. 1980. Introduction to the Tiberian Masorah. Trans-
lated and edited by E. J. Revell. Masoretic Studies 5. Mis-
soula, MT: Scholars Press.

. 2003. The Biblical Masorah. Jerusalem: The Hebrew Acad-

emy of Language. [Hebrew]

TIBERIAN KETIV-QERE AND THE
COMBINED SAMARITAN WRITTEN-
READING TRADITION: POINTS OF
CONTACT AND CONTRAST

Aaron D. Hornkohl

The phenomenon of ketiv-gere is the clearest indication of the
composite nature of the Tiberian biblical tradition. Against the
backdrop of the normally harmonious relationship between the
tradition’s written (i.e., consonantal, orthographic) and pronun-
ciation (i.e., vocalisation, recitation) components, such acknowl-
edged cases of written-reading dissonance are clear evidence of
divergence (see Khan 2013a; 2020, I:33-49).

Crucially, however, beyond this, the phenomenon is
opaque. The ketiv-gere mechanism signals, but does not explain,
discord within the tradition, which is left for scholars to illumi-
nate. It is sometimes assumed that the pronunciation tradition
‘protects’ or ‘corrects’ readings that have become garbled in the
written tradition. While this may occasionally be the case (espe-
cially in cases of possible conflation of waw and yod, relevant in
more than one case below), the view that it is the norm fails to
do justice to the relationship between the ketiv and the gere. In

many cases, both represent plausible readings. It is thus simplest

116 Hornkohl

and most appropriate to think of the two components as trans-
mitted artifacts that represent related but distinct traditions of
pronunciation and interpretation—a major difference being that
the written component only partially and ambiguously reflects
how it was ever orally realised, whereas the pronunciation com-
ponent does so more comprehensively and precisely. In sum,
throughout the vast majority of the biblical text, there is no evi-
dence to suggest anything other than harmony between the two
components of the tradition; but in a not insignificant minority,
the two clearly diverge.

In a number of cases in which the ketiv and gqere offer syn-
onymous linguistic alternatives, the gere reflects the characteris-
tically later option. This is consistent with the view that the oral
development of the Tiberian reading tradition, which was ulti-
mately recorded in the vowel signs superimposed on medieval
consonantal manuscripts, was largely complete by the end of Sec-
ond Temple times. Having crystallised in the late antique period,
it was something of a mixed linguistic system, regularly preserv-
ing features of more ancient Hebrew and simultaneously incor-
porating later secondary features. It should come as no surprise,
then, that instances in which the reading component diverges
from its written counterpart—whether or not explicitly acknowl-
edged in masoretic sources—often show signs of secondary lin-
guistic development (Khan 2013b; 2020, 1:56-85; Hornkohl
2018, 86-91; 2020a, 248-57, 263-64; 2020b, 420-22).

For its part, the Samaritan biblical tradition is also compo-
site, comprising related but independent written and recitation

components that blend First and Second Temple traits. Letter

Tiberian ketiv-qere and the Samaritan Tradition 117

shapes, word-separation dots, content, and much in the way of
linguistic data hark back to the Iron Age, suggesting antiquity.
The orthography and some minority linguistic features, on the
other hand, display much in common with Hebrew and Aramaic
sources that date from the end of the Second Temple period (Ben-
Hayyim 2000, 3-4, 80.4; Tal and Florentin 2010, 25-28). This
tallies with Tov’s (2012, 79) summary, which emphasises that
many features considered distinctively or especially Samaritan
already distinguish proto-Samaritan manuscripts from proto-
Masoretic material at Qumran, indicating that “the s.-group re-
flects a popular textual tradition of the Torah that circulated in
ancient Israel in the last centuries BCE, in addition to the 2%-group
and other texts.” Some very late developments in the Samaritan
reading tradition (see, e.g., the Samaritan phonology in §3.0 be-
low) are due to contact with Arabic (see, e.g., Ben-Hayyim 2000,
29, §1.0.1, 32-33, §1.1.4).

There is no exact Samaritan counterpart to the Tiberian
ketiv-gere mechanism. Even so, due in part to the chronological
distance between the respective linguistic traditions embodied in
the written and reading components of the Samaritan Penta-
teuch, dissonance between the two is commonplace. It is far more
frequent than in the combined Tiberian written-reading tradi-
tion, with the Samaritan recitation tradition regularly ‘updating’
the oral realisation ostensibly reflected in the consonantal text—
which, to be sure, itself shows occasional contemporisations in
accord with Second Temple conventions (Hornkohl 2021, 8-9;

see also 811.0, below). Additionally, it is also important to bear

118 Hornkohl

in mind the Samaritan biblical tradition’s penchant for harmoni-
sation, a characteristic that extends from the ironing out of per-
ceived discrepancies in content to grammatical levelling and the
imposition of morphosemantic order (Ben-Hayyim 2000, 121-
22, §2.2.1.1.5; Tal and Florentin 2010, 28-34; Tov 2012, 80-86;
Hornkohl 2021).

The present study examines a selection of verb-centred
ketiv-gere instances in the Tiberian Pentateuch (based primarily
on L, with comparison to A where possible and appropriate).' In
the following discussions, an attempt is made to explain cases of
Masoretic ketiv-gere dissonance, to compare the relevant Samari-
tan written form and oral realisation, and to contextualise all

within broader historical linguistic trends.

1.0. Qere xyrn || ketiv xxin (Gen. 8.17)

Throughout the MT, hifil s*yin ‘bring out, take out’ presents the

following imperatival forms:

11 = first person; 2 = second person; 3 = third person; A= Aleppo
Codex; BA = Biblical Aramaic; BH = Biblical Hebrew; c = common
(gender); CBH = Classical Biblical Hebrew; CGT = Cairo Geniza Tar-
gum; DL = dual; F = feminine; FT = Fragment Targums; K = ketiv; L
= Leningrad Codex (Firkovich B 19 A); M = masculine; MT = Maso-
retic Textual Tradition; PL = plural; Q = gere; QA = Qumran Aramaic;
QH = Qumran RH = Rabbinic Hebrew; s = singular; SAT = Samaritan
Arabic Translation; SP = Samaritan Pentateuch; ST = Samaritan Tar-
gum; t. = Tosefta; TN = Targum Neofiti; TO = Targum Onkelos; tr. =
transitive; TY = Targum Yerushalmi (i.e., Targum Pseudo-Jonathan);
y. = Talmud Yerushalmi.

Tiberian ketiv-qere and the Samaritan Tradition 119

MS: 8¥in/-X’vin (Gen. 8.17 ketiv; 19.5, 12; Exod. 3.10; Lev.
24.14; Judg. 6.30; 19.22; 1 Kgs 22.34; 2 Kgs 10.22; Isa.
43.8; Ezek. 24.6; Ps. 25.17; 142.8)

FS: "iT (Josh. 2.3)

MPL: IN¥IN/-ANYIA/-NVIT (Gen. 38.24; 45.1; Exod. 6.26; Josh.
6.22; 10.22; 2 Sam. 13.9; 1 Kgs 21.10; 2 Kgs 11.15; Isa.
48.20; 2 Chron. 23.14; 29.5)

The lone exception is gere xx7n || ketiv xxin in (1).

(1) oy wAdD wAIT OR APTA Ava Wale TARTWs AnD
{PINT OW AIT TID) PISA WW TAS (Q) NBN (K) 8317 PINT

‘Every living thing that is with you of all flesh—birds and
animals and every creeping thing that creeps on the earth—
bring out with you, that they may swarm in the earth, and

be fruitful and multiply on the earth.’ (Gen. 8.17)

Depending on one’s expectations of the ketiv-gere phenomenon,
this instance may be surprising. It is not uncommon in cases of
written-reading dissonance for the gere to reflect a usage more
conventional than that reflected in the ketiv. Here, however, the
situation is reversed. Ketiv xin matches the form that occurs in
the 24 other occurrences of this verb’s imperative, whereas the
qere is unique.

A well-known Hebrew feature is the merger of original I-w
and I-y verbs, especially the shift from I-w to I-y in syllable-initial
position, e.g., gal 17 ‘give birth’ (cf. Arabic J,; Blau 2010, 104,
§3.4.8.6-9, 245-46, 84.3.8.4.4-8n). The original w seems to have
fared better in other environments, e.g., as the offglide of a diph-
thong, but even there it frequently loses consonantal force due to

monophthongisation, e.g., nif‘al 1911 < *nawlad ‘be born’ (Blau

120 Hornkohl

2010, 228, §4.3.5.3.2), hifal toin < *hawlid ‘beget’ (Blau 2010,
235-36, §4.3.5.7.5). The shift I-y to I-w, restricted chiefly to non-
word-initial y, is also known, e.g., wa ‘be/become dry’ (cf. Arabic
un), but hifil wrin “dry (tr.)’. Of special relevance in this con-
nection is the case of ketiv win gere wn ‘make level! (Ms) (Ps.
5.9). In this ketiv-gere instance, it would seem that the orthogra-
phy reflects a tradition in which an original I-y form was realised
as if it were I-w due to analogical pressure of the majority shift y
> w (as sometimes in Aramaic, in the case of this root; see CAL,
s.v.). The gere, conversely, is in accord with the conventional I-y
hifil form as evidenced elsewhere in the combined Tiberian writ-
ten-reading tradition as well as in other ancient Hebrew sources
(including the DSS and Ben Sira).

In the case of the root 8"x"/1, on the other hand, the Semitic
evidence (e.g., Ethiopic wad’a, Old South Arabic wd? or wz) seems
to indicate an original I-w form, which secondarily shifted to I-y in
Northwest Semitic (Hebrew, Aramaic, Phoenician, Ugaritic).

As the expected form, the ketiv requires no explanation. For
the qgere there are various explanations. It may reflect truly an-
cient phonological diversity that was generally levelled in favour
of the dominant y > w shift (cf. the related Aramaic shaf‘el *yw,
which also preserves the y). Along these lines, Cohen (2007, 53-
54) has suggested that the reading tradition exploited the option
of an exceptional I-y form in Gen. 8.17 in the interests of aural
euphonic repetition, xx’ ‘bring out’ (v. 17), -xx ‘and (Noah)
came out’ (v. 18), ix ‘came out (MPL)’ (v. 19). Alternatively, it
is not impossible that the gere here stems from a written tradition

in which an ambiguous waw was misinterpreted as a yod. This

Tiberian ketiv-qere and the Samaritan Tradition 121

would suggest that at least some cases of gere might stem from
the reading of manuscripts, rather constituting a purely oral en-
deavour.’

For its part, the combined written-reading Samaritan tradi-
tion at Gen. 8.17 has x°xin usi. The long u-vowel is standard in
Samaritan I-w hifil verbs (as the open-syllable equivalent of
closed-syllable short 0; Ben-Hayyim 2000, 44, §1.2.0), as is the i-
vowel in the open second syllable of verbs III-’. Given the SP’s
penchant for levelling and harmonisation, its presentation of a
standard imperative here is not unexpected, though in this case
it also occasions the rather rare agreement of the Samaritan tra-
dition with the Tiberian ketiv.

2.0. Qere nvr || ketiv ow (Gen. 24.33)

An acknowledged feature of late antique Hebrew involves shifts
of G- to C-stem, i.e., gal to hifil, with no accompanying semantic
change. The shift appears to have been especially frequent in,
though by no means exclusive to, hollow, i.e., II-w/y, verbs, e.g.,
derivations of 7'""1 ‘understand’, 1" ‘act arrogantly’, x""p ‘vomit’,
a" ‘quarrel’, p"*> ‘scoff (Hornkohl f.c.).

While the Tiberian reading tradition is opaque with regard
to the analysis of II-w/y prefix conjugation (yiqtol) verbal forms,
i.e., whether they are qal or hifil (as opposed to suffix conjuga-
tion [gatal] forms, participles, and infinitives), this is not the case
with hof‘al forms. Based on regular sound changes, the expected

qal passive wayyigtol form of the verb ow ‘put’ is nw) ‘and it was

? Tam grateful to Geoffrey Khan for raising this possibility.

122 Hornkohl

put’, on which the Tiberian written and reading traditions agree

at Gen. 50.26; see example (2).°

(2) oyna xa OY) ink inn ow IwpI ANATTA NeP nan
‘So Joseph died, being 110 years old. They embalmed him,
and he was put in a coffin in Egypt.’ (Gen. 50.26)

This is precisely the orthography one finds in the ketiv nw (Gen.
24.33), but the corresponding gere nw) ‘and it was put’ is hof‘al
(Cohen 2007, 63-64; Blau 2010, 97, §3.4.3.3; cf. GKC 148, 284;
Bergstrasser 1918-1929, I:459); see example (3).

(3) Mat omat ox Ty Dak XD dosh 54x Pia) (Q) OWI) (K) ow

SAT TAN
‘Then food was set before him to eat. But he said, “I will
not eat until I have said what I have to say.” He said, “Speak
on.” (Gen. 24.33)

This evidently reflects three related secondary developments: (1)
passive formation of II-w/y verbs on the analogy of I-w/y, (2) the
well-known decline of the gal internal passive, and, since hof«al
represents the internal passive of hifil, (3) hifilisation, i.e., the
broad movement from qal to hif‘il with no corresponding semantic
shift (Blau 2010, 97, §3.4.3.3). In other words, a realisation such
as gere DW" implies the existence of hif‘il nw7, as seen occasionally
in the Tiberian written tradition (Ezek. 14.8; 21.21; Job 4.20) and
more commonly in late antique extra-biblical Hebrew (SirA 4v.22
= Sir. 11.30; t. Gittin 7.13; Sifre Devarim 315; y. Sanhedrin 1.1;
frequently in the Babylonian Talmud).

3 According to Blau (2010, 97, §3.4.3.3), the expected resolution of the
diphthong uy is contraction to i, thus wayyigem < *wayytysem.

Tiberian ketiv-qere and the Samaritan Tradition 123

The Samaritan form that corresponds to Tiberian ketiv ow
gere nw (Gen. 24.33), as well as to Tiberian nw1 (Gen. 50.26),
is ow" wyuwwdsam. The written form might conceivably reflect
the same qal internal passive > hof‘al shift as seen above in the
Tiberian ketiv-qere. Indeed, Ben-Hayyim (1977, 271) formally
classifies the form as hifil passive. Crucially, however, the form
realised in the Samaritan pronunciation tradition does not reflect
the hof‘al stem, i.e., the internal passive of hif‘l, but an external
passive, in this case most probably Gt, with assimilation of the
infix -t-, in the following manner: yuwwasem < yiwwasem < yit-
wasem (alternatively, Dt, with assimilation of the infix -t- and
simplification of middle-radical gemination; Ben-Hayyim 2000,
178, §2.10.4). And, of course, this may well underlie the Samar-
itan written form, as well. Thus, like the Tiberian reading tradi-
tion in the ketiv-gere in Gen. 24.33, the Samaritan reading tradi-
tion replaces the archaic qal internal passive with a secondary
and more contemporary alternative. While the Tiberian gere tal-
lies with rather common hof‘al use throughout the combined Ti-
berian written-reading tradition, the Samaritan recourse to Gt
forms, especially with assimilated -t- (or to Dt forms with assim-
ilated -t-) smacks of late Aramaic linguistic practices uncharac-

teristic of the early Hebrew linguistic sources.

3.0. Qere nnnw || ketiv innwn (Gen. 27.29; 43.28)

Twice in the MT, ketiv-gere discord focuses on yiqtol forms of the

verb minnwn ‘bow down’; see examples (4) and (5).

124 Hornkohl

(4) ord Sax min ox 7 (Q) WW (K) INN oy qT

SF] PIII WS PTR TWaR ya 77 Ww
‘Let peoples serve you, and nations bow down to you. Be
lord over your brothers, and may your mother’s sons bow
down to you. Cursed be everyone who curses you, and
blessed be everyone who blesses you!’ (Gen. 27.29)
(5) (Q SAHA (K) INN sappy wae? TTA? ow TAN
‘They said, “Your servant our father is well; he is still alive.”
And they bowed their heads and prostrated themselves.’
(Gen. 43.28)

In Tiberian Hebrew, the 3Ms prefix conjugation form of this verb
is unique, in that it ends with -i, innv> ‘he bowed down’, cf. MPL
ynnw ‘they bowed down’.* Since the ketiv form resembles the
relevant 3MS yiqtol form, it may at first glance be tempting to
argue that the ketiv simply construes as singular what the qere
construes as plural. Given the context in both cases, however, this
seems unlikely. Both passages include other clear instances of
3MPL yiqtol forms in proximity, including, in example (4), explic-
itly 3MPL 1iNMW later in the verse.

If the solution is not morphosemantic, perhaps it is phono-
logical. 3mMs innmv yistahi and the reconstructed precursor of

3MPL Ninny yistahdwi < *yistahwii are distinguished by only the

* The unique form would appear to be a natural consequence of the
syllable structure of short a prefix conjugation (yigtol) hishtaf‘el form
from root "in/1"1n, yistahai < yistahu < yistahw, in which the vowelless
radical w of the word-final consonant cluster hw could be preserved only
as a long u-vowel (for analysis as a derivation of 1"nw see Blau 2010,
237, §4.3.6.1).

Tiberian ketiv-qere and the Samaritan Tradition 125

onglide in the diphthong -wi. In the absence of an epenthetic
hatef vowel to resolve the -hw- consonant cluster, the w would
have been extremely vulnerable to syncope and, presumably,
graphic non-representation.

Another possibility, raised by Cohen (2007, 18), is that in
these cases, as well as those of ketiv 1x» || gere 1iv71 ‘and they com-
manded’ (Jdg. 21.20), ketiv 7nx7 || gere AN" ‘and they said’ (1
Sam. 12.10), and ketiv 117% || gere Mam ‘and they spoke’ (1 Kgs
12.7), the ketiv represents rare defective spelling of the word-final
plural morpheme.

And, of course, in all of the above cases there is the possi-
bility of simple textual corruption, i.e., the accidental graphic
omission of the expected waw.

Whatever the most compelling explanation for the relevant
Tiberian ketiv-gere, the SP shows no trace of dissonance. Both the
written component of the tradition and its reading counterpart
reflect standard 3MPL forms: 1nnw™ wyistabbu. The spelling re-
veals no disharmony between these plural forms and others in
the vicinity nor between these plural forms and other plural
forms of this verb.

The Samaritan pronunciation deserves special comment.
Evidence indicates that the early realisation of Samaritan waw,
namely w, shifted to v in the Second Temple period and that later,
due to coalescence of v < w and v < b, most cases of v (< w)
were included in the general b < v shift due to Arabic. The dou-
bled middle radical may, as in Tiberian Hebrew, reflect pattern

gemination, but it is also possible that it derives from regular as-

126 Hornkohl

similation of the guttural h, i.e., -bb- < -ww- < -w- < -hw-. Sa-
maritan Hebrew also more conspicuously distinguishes between
singular \nnw” yistdbbi and plural wnnw” yistdbbu, the singular re-
alised as a standard, rather than short, III-y prefix conjugation
(yiqtol) form. It should be noted that all of the above develop-
ments in the Samaritan tradition are secondary features that re-
flect phenomena that typologically post-date the form of the Ti-

berian gere.

4.0. Qere 73 8a || ketiv 122 (Gen. 30.11)

(6) ta inw-ny NPM (Q) TW NB (K) 733 AX? wAXM
‘And Leah said, “Good fortune has come!” so she called
his name Gad.’ (Gen. 30.11)

Though it has been suggested that the Tiberian ketiv and gere in
(6) are mere phonological variants reflecting a single common
exegetical tradition (Cohen 2007, 42-43), the testimony of an-
cient witnesses arguably indicates otherwise, i.e., that they re-
flect diverging interpretations. Ketiv 131 is taken as an adverbial
in the sense of ‘with good fortune’; cf. LXX ’Ev tuyy ‘with luck’;
Vulgate feliciter ‘happily’. Qere 73 8a,° on the other hand, is a verb-
subject verbal clause; cf. Peshitta x. «ur ‘fortune has come’;
TO 73 xnx ‘fortune has come’; TY x20 x51 NN ‘good luck has
come’; TN/FT/CGT x20 773 Xnx ‘good fortune has come’. The

ketiv and gere variants are of approximately equal plausibility and

° In L, the marginal gere notation is especially detailed, including not just
the conventional information of consonants and word separation (with
x2 and 73 on separate lines), but also vocalisation (minus the dagesh in
the bet of x1 and the rafe over the gimel of 73) and accentuation.

Tiberian ketiv-qere and the Samaritan Tradition 127

each enjoys support, both ancient and medieval (on the latter,
see Habib 2020, 318).°

The combined Samaritan written-reading tradition has 731
afgad, which the ST renders 7101 ‘tidings, news’, apparently in
line with the Tiberian gere. Cf. the SAT, which renders Sus t+
‘an army has come’, which rather corresponds to the Tiberian
ketiv. Despite the diachronic proximity of the Tiberian gere (and
reading tradition, more generally) and the Samaritan reading tra-
dition, the latter sometimes agrees with the Tiberian ketiv. This

appears to be such a case.

5.0. Qere isi || ketiv v5 (Exod. 16.2); gere
won Ary || ketiv inon/in'’m (Exod. 16.7; Num.
14.36; 16.11)

In Tiberian Hebrew, the root 7"1%/7'"% II ‘grumble, complain’ is
represented by largely synonymous nif‘al and hif‘il forms. Beyond
the written-reading deviation at issue here (which is not neces-
sarily limited to the acknowledged instances of ketiv-qere), sev-
eral additional factors combine to complicate the Tiberian para-
digm of "15/7" II: (a) potential conflation of II-w and II-y forms
(at both linguistic and textual levels); (b) partial homophony
with forms of 7"15/{""5 I ‘lodge, spend the night’ (against which
problem, secondary morphological gemination developed in
some forms of {"15/;""5 II; (c) the morphosemantic challenge of the

formal distinction between intransitive, transitive, and causative

°In L, a note in the bottom margin of the page including (6) (fol. 17v)
lists cases of single-word ketiv versus two-word gere and vice versa.

128 Hornkohl

senses; (d) broad morphosemantic movement away from qal in
favour of morphology perceived to have greater semantic iconic-
ity.

As reflected in the Tiberian reading tradition, the fourteen
occurrences of 7'"15/1'"" II ‘grumble, complain’ seem to comprise a

suppletive paradigm with notable outliers. See Table 1.

Table 1: Tiberian forms of "1b/;""5 II by TAM form and stem (binyan)

TAM form Case (reference) Stem (binyan)

pop op oAN wR o>nidn (Exod. 16.8)
gatal forms orn (Num. 14.27a)
and ~y opbn nag Ws Ok ya nifyn (Num. 14.27b)
participles oniryn (Num. 14.29)
oxy ordn on ws Sky ra hixyn (Num. 17.20) hifil

avon (Exod. 16.7 gere; ketiv indn)

tym (Exod. 17.3)

aon (Num. 14.36 gere; ketiv 1319)

avyn (Num. 16.11 qere; ketiv indn)

am (Exod. 15.24)

sit) (Exod. 16.2 qere; ketiv ir")

i351 (Num. 14.2) nif‘al
335" (Num. 17.6)

ii (Josh. 9.18)

(way)yiqtol
forms

The lone suffix conjugation form (Num. 14.29) and all participles
(Exod. 16.8; Num. 14.27a, 27b; 17.20) are consonantally unam-
biguous hifil forms. It may, however, be significant that three of
these forms—and no others—are explicitly transitive, taking as
direct object a form of 739m ‘complaint’ (Exod. 16.8; Num.
14.27b; 17.20), while the remaining two forms (Num. 14.27a,
29) occur in the same context. The lone causative prefix conju-

gation form (Num. 14.36) is also hifal (and, critically, one of the

Tiberian ketiv-qere and the Samaritan Tradition 129

cases of ketiv-gere discord). Of the eight remaining prefix conju-
gation forms, three are hifil (Exod. 16.7 gere; 17.3; Num. 16.11
qgere; two involve ketiv-qere dissonance), and five are nif‘al (Exod.
15.24; 16.2 gere; Num. 14.2; 17.6; Josh. 9.18; one involves ketiv-
qere dissonance). It may be significant that all yigtol forms in the
reading tradition are hifil, whereas, with two notable exceptions,
wayyiqtol forms are nif‘al: the exceptions are 171 (Exod. 17.3) and
the causative gere ird1 (Num. 14.36) (on both of which see be-

low). For another perspective consider Table 2.

Table 2: Tiberian forms of }"1b/7'""5 II in canonical order

Reference Form in context Stem Semantics Source
Exod. 15.24 awirdy op 179 N intr. J
Exod. 16.2Q pax oy awirdy wea nba id ON intr.

2K Aap wirdy yea mb. Po" ?

7Q ardpardnes H tr.?

7K ardpandnes N ?

8 roy OF 5N onx-wy od nda H tr. P
Exod.17.3 nwirdy ova 72" H ? E
Num. 14.2 Ssnv a ba aad away ID" N intr P

27a Sy ord naa H tr.? P

27b sy Oo naa Wes dsqw a nifbn H tr. P

29 pont awe H tr.? P

36Q raya any poy ary" H caus

36K mapa barny pop nd" N intr g
Num. 16.11Q roy arn H ?

11K rwindns N ?
Num.17.6 9 yy awi by... mba 591 N intr. P

20 oxy on on rwe dow a2 hisbn H tr. P
Josh. 9.18 mya dy mado 5" N intr.

It is difficult to conceive of an exhaustively satisfying account of
the particular constellation of forms as reflected in either the or-
thographic or the recitation tradition, including the four
acknowledge ketiv-qgere cases, since a comprehensive mor-

phosemantic rationale for the use of hif‘l versus nif‘al (way)yiqtol

130 Hornkohl

forms is elusive. Neither does recourse to putative Pentateuchal
source provide clarification. It is, of course, possible that graphic
confusion between waw and yod is relevant in some cases. It may
also be that the hifi and nif‘al forms are, at least to some extent,
synonymous and grammatically interchangeable—though distri-
butional differences—especially the exclusive use of hif‘l for un-
equivocally transitive cases—militate against this. Another factor
worthy of consideration is contextual proximity. Note that in
three pericopes with multiple forms, i.e., Exod. 16, Num. 14, and
Num. 17, the initial grumbling is indicated via an apparently in-
transitive nif‘al, whereas the use of hif‘il forms ensues only in the
immediate vicinity of another explicitly transitive hifil (Exod.
16.8; Num. 14.27b; Num. 17.20). It seems reasonable to postu-
late that these forms were realised as hif‘il, whether due to attrac-
tion or to analysis as genuine, if elliptical, transitives (noted by
‘tr.?’ in Table 2). Wayyiqtol forms are generally nif‘al, unless caus-
ative—though 19" (Exod. 17.3) remains an outlier.’

Moving to the specific cases of ketiv-qere dissonance,
graphic and/or linguistic conflation of II-w and II-y may be ap-
plicable in any or all cases (Cohen 2007, 72-73).° Without defin-
itively ruling out these possibilities, the following discussions will

consider alternative hypotheses.

” Though pure conjecture, it is possible that the apparently hif‘l 17 is
in reality an old intransitive qal II-y form. While other intransitives were
realised as nif‘al, this case may have retained its realisation due to sim-
ilarity with hif‘l forms.

8 Textually, the graphic similarity—or, in some cases, identity—between
waw and yod requires no elaboration. On the linguistic level, consider the

Tiberian ketiv-qere and the Samaritan Tradition 131

Perhaps the most straightforward case involves the causa-
tive in (7).

(7) (Q) 39791 (K) NS saw payTrns and Awa Mow wY OT

mpa ane Pop

‘And the men whom Moses sent to spy out the land—they

returned, and they made all the congregation grumble
against him...’ (Num. 14.36)

Here the hif‘il form seems especially fitting for the double-transi-
tive causative semantics. The ketiv may reflect local exegesis dif-
ferent from that represented by the gere, according to which the
words ntya-a-nx Py indy were understood to mean not what
the returning spies did to the people, but what they did with the
people: ‘and they grumbled with the congregation’ (rather than
‘and they made the congregation grumble’).’

Turning to example (8):

(8) aTAa AAR yp AWAMDY Dene ia nIy-5> (Q) aide (K) roy
‘And the whole congregation of the people of Israel grum-
bled against Moses and Aaron...’ (Exod. 16.2)

If not a simple corruption or synonymous linguistic variant, the
hifil ketiv morphology in example (8) is consistent with the
nearby hifil participle at Exod. 16.7—though, admittedly, at
odds with the nif‘al ketiv morphology in the neighbouring verse

regular pairing of gal yigtol n°w and infinitival Diw(%) ‘put’ and of qal yigtol
r? and infinitival 1355 ‘lodge’. Further examples could be adduced.

° Cf. LXX dteydyyvoav xat adtijs mpds thy cuvaywyny ‘(they) murmured
against it [i.e., the land] to the assembly’, in which, to be sure, the sense
is neither causative nor comitative.

132 Hornkohl

in example (9), below. The nif‘al gere is consistent both with
other nif‘al intransitive forms—contrasting with transitive or
causative hifil morphology—and with the nif‘al majority of way-
yiqtol forms.

Examples (9) and (10) reflect the same ketiv-gere disso-

nance in very similar usages.

(9) srby (Q) ron (K) NOM 19 nA ann...
‘...For what are we, that you grumble against us?’ (Exod.
16.7)"°

(10) :roy (Q) won (K) n9n +2 gxa-n aN...
‘,..And Aaron, what is he that you grumble against him?’
(Num. 16.11)

Due to the verbal similarity, i.e., near parallel structures compris-
ing by + 1v/bn + 72, it is no surprise that the yigtol forms have
matching stems within the Tiberian reading and written tradi-
tions, respectively. There seems to be logic to both alternatives.
The hifil gere in (9) is consistent with the explicitly transitive
hifil participle in the following verse. Conceivably, the hif‘l gere

in (10) was also deemed transitive, or was simply read as hif‘il,

10 In L, the ketiv-gere instance at Exod. 16.7 is signalled by means of a
circellus above the ketiv, which is vocalised and accented according to the
reading ir9n, as well as by an intercolumnal notation reading 7p *. Upon
close inspection, however, L’s ketiv does not unambiguously read indn.
The ostensible first waw is noticeably shorter than the second and is more
similar in shape to the yod in the next word, i9y. Cf. the waws and yods
in the surrounding context. It is possible that the gere orthography has
actually found its way into the written tradition here, though it has been
furnished with a ketiv-gere note consistent with the masora.

Tiberian ketiv-qere and the Samaritan Tradition 133

on the basis of its similarity to (9). At any rate, according to the
qere, all forms except wayyiqtol are hifil. As for the ketiv—it is
conceivable that the influence worked in the opposite direction,
i.e., the yigtol form in example (10) was deemed an intransitive
nif‘al and verbal similarity determined the nif‘al realisation of the
near parallel instance in (9), all of which resulted in consistently
nif‘al yiqtol forms in the orthographical tradition, contrasting
with hif“il-nif‘al diversity in the case of wayyiqtol forms.

Having attempted to clarify the complex situation in the
Tiberian tradition, we may turn to the rather simpler situation in
the SP. Here, as in Tiberian Hebrew, gatal and participial forms
are hifil, e.g., oni5n allentimma and 0°15n mallénam, respectively.
Unlike in the MT, however, all prefix conjugation forms—
whether yiqtol or wayyiqtol, and no matter their semantics—are
gal, e.g., to" wyilldn, 15" wyilldnu, 1n tillanu. Whether gal or
hifil, forms consistently reflect geminate analysis (i.e., 1"15),
which in Samaritan Hebrew routinely involves gemination of the
first radical, on the I-n pattern (see Ben-Hayyim 1977, 154; 2000,
156, 82.7.6). While the Samaritan derivation and stem arrange-
ment show no morphological distinction between intransitive
and causative semantics,'! thanks to the geminate derivation,
there is no chance of homophony with forms of 7"1>/7'""5 I ‘lodge,
spend the night’, which—whether qal or hif‘il—in Samaritan He-
brew has the form of a hifil with gemination of the first radical
bn wyallan, iron wyallinu, pon tallan (Ben-Hayyim 2000, 152,
§2.6.13).

" Against the SP’s equivalence of intransitive and causative forms, the
ST’s syntax in Anw39 52 mn’ yyy 11 reveals a causative reading of the verb.

134 Hornkohl

Two further points seem relevant. First, it is noteworthy
that both the Tiberian and Samaritan recitation traditions exhibit
what must be considered secondary gemination in the case of
forms of 7"1b/7'""5 II ‘grumble, complain’. Though the explanations
for gemination in each tradition differ—lexeme-specific semantic
disambiguation in Tiberian (Yeivin 1980, 362; Khan 2020, 1:524)
and broader paradigmatic pattern suppletion in Samaritan—the
mere fact of the shared trait arguably points to its early, pre-
schism development.

Second, there is the matter of stem morphology in both tra-
ditions. Given the dissonance and uncertainties discussed above,
the antiquity of the hif‘l-nif‘al Tiberian arrangement may be
questioned. The Samaritan hif‘il-qal arrangement adds to the un-
certainty. Is it possible that the Tiberian nif‘al goes back to an
earlier gal, which was preserved in the Samaritan tradition? The
secondary nifalisation of original gal verbs with intransitive and
middle semantics is a feature of the evolution of ancient Hebrew
as seen in the extant sources, especially relevant to Second Tem-
ple chronolects, e.g., the Tiberian LBH orthographical tradition,
the Tiberian biblical recitation tradition more broadly, and the
Samaritan biblical tradition (Hornkohl 2021). Regarding the lat-
ter, it should be borne in mind that, (a) Samaritan 7"s>/;"" II
‘grumble, complain’ is analysed as a geminate verb; (b) Samari-
tan geminate verbs are routinely realised as I-n forms with assim-
ilated nun; (c) Samaritan I-n forms with assimilated nun (and sim-

ilarly realised geminate forms) are ineligible for nif‘al analysis;'?

preservation of gal internal passive np” (Exod. 21.21), both ‘he must be

Tiberian ketiv-qere and the Samaritan Tradition 135

and (d) the Samaritan qal vowel pattern is that of the dominant
yiqtal template rather than that expected of II-w/y forms. Thus, a
speculative, though not implausible hypothesis is that an original
qal underwent nifalisation in Tiberian Hebrew but was preserved,
albeit with secondary gemination and vocalism, in Samaritan He-
brew (see Hornkohl 2021, 9-10).

6.0. Qere *xinp || ketiv xp (Num. 1.16); gere *“"IP
|| ketiv xp (Num. 26.9)

The cases of ketiv-gere dissonance in (11) and (12) are mirror im-

ages.'®

(11) 7-998 (WRT ONIN niga ~yyI TIA (Q) NIT (K) (NTP 7x

DF RW?

‘These were the ones chosen from the congregation, the

chiefs of their ancestral tribes, the heads of the clans of Is-
rael.’ (Num. 1.16)

(12) (QU)°8TP (K) NTP OP agT IT Na OPIN] TNT) FRI AWN

maby onkaa Mp nwa AX Sy) awh dy WA WEY TVA

‘The sons of Eliab: Nemuel, Dathan, and Abiram. These are

the Dathan and Abiram, chosen from the congregation,

who contended against Moses and Aaron in the company

of Korah, when they contended against the LorpD’ (Num.

26.9)

avenged’. Ineligible for nif‘al analysis and realisation, the latter retained
its gal passive vocalism, though it may also have been identified as
hof«al.

'3 In L, a masoretic note at the bottom of the page that includes (11) (fol.
74r) lists ketiv-gere instances involving interchanges of waw and yod.

136 Hornkohl

Both involve substantives related to the verb x4) ‘call, read’, in
the qdtil and qatil nominal patterns. The former is common, but
not systematically productive in ancient Hebrew “for stative or
passive actant nouns, mostly adjectives, but also with secondary
substantive meaning, especially for the passive ones” (Fox 2003,
192; Huehnergard 2007). For its part, qdtil in ancient Hebrew “is
a completely productive patiens participle, serving for the object
of transitive verbs” (Fox 2003, 201). While no historical phase of
Hebrew lacks qatil or related gatil forms, specific lexemes are lim-
ited to, or especially characteristic of, late Hebrew chronolects.
This is possibly due in part to contact with Aramaic, in which the
related qatil template is fully productive as the G-stem passive
participle (Fox 2003, 195). Examples include: pni ‘temple serv-
ant’ (LBH; QH; RH; Samaritan Hebrew), n*7w ‘messenger’ (RH; cf.
BA; QA; TA), pon ‘strong’ (1x in BH; various Aramaic dialects),
and 7°78 ‘official’ (CBH [rare]; LBH; QH; Samaritan Hebrew; cf.
Egyptian Aramaic).

In (11) and (12) above, the two terms have the general
sense of ‘leaders’. Some have sought a semantic distinction be-
tween 77vA (NP and ATYA NTP, but Cohen (2007, 197-98) ad-
duces arguments and references against such an approach. Sig-
nificantly, the ketiv in one place confirms the validity of the qgere
in the other and vice versa. Additionally, each form is also found
in a similar context elsewhere in biblical literature, where the
written and reading components of the tradition apparently co-
incide: cf. TyIA NIP ATV °N’w:I ‘chiefs of the congregation, chosen
from the assembly’ (Num. 16.2) and 033 03303 nind Jan “na

DxIPI Dwow ‘desirable young men, governors and commanders

Tiberian ketiv-qere and the Samaritan Tradition 137

all of them, officers and men of renown’ (Ezek. 23.23). Each of
the components of the combined Tiberian tradition seems to bear
witness to a situation of genuine lexical diversity, differing only
with respect to the proper context for the respective forms."

In the SP, the forms in examples (11) and (12) are both
written and read °xp qgary@i (< *qarydy < “*qariyydy <
*qarvay, from the gqatil template + Aramaic gentilic -dy; Ben-
Hayyim 2000, 284, 84.3.8). This is consistent with the form in SP
Num. 16.2, where the MT has °x7. That the Samaritan tradition
should unify forms, and do so in favour of what constitutes the
majority form (in Samaritan as well as Tiberian), is not surpris-
ing. But there may be more to the story, with the choice of forms
being part of a broader preference for late forms characteristic of
Aramaic and RH. Ben-Hayyim (2000, 199-200, §§2.13.2-4) lists
qatol, gétal, and qdtal as templates for the Samaritan qal passive
participle, all of which bear marks of post-classical development.
The shift from PS gatiil to qdtol can be explained in line with reg-
ular Samaritan vowel changes, but based on qdtol’s use for the
qal active participle, one may infer the influence of the qdtol no-
men agentis pattern so common in late Aramaic, especially Syriac,
and RH (Hornkohl 2013, 148-52). For their part, the Samaritan
gétal and qdtal templates may both have developed from PS qatil,

Consider the interchange between the approximately synonymous
English venerable and venerated in “The specimens of the venerated
Bede, as given by Colonel Dow before his History of Hindustan, exhibit
rhyme” (Morgangw 1858, 354) and the more customary appellation,
reflecting Catholic soteriology, “the Venerable Bede.”

138 Hornkohl

but alternatively reflect Aramaic gatil and Hebrew qatil, respec-

tively.’

7.0. Qere win || ketiv wm (Num. 21.32)

In Tiberian BH there is a general distinction between qal w7 and
hifil win, in that the gal typically takes an inanimate object and
means ‘inherit, take possession of, whereas the hif@l tends to take
an animate object and to denote the sense of ‘dispossess, disinherit,
drive out; cause to inherit’. Not infrequently, however, there is se-

mantic and grammatical reversal (see the standard lexicons).

15 Whatever the case, the incidence in the SP of qdtal and gqétal for the
G-stem passive participle is comparatively greater than in the Tiberian
Torah, including: nya bildt ‘married (to a husband)’ || MT n>pa (Gen.
20.3; Deut. 22.22), on égiram ‘girded’ || MT o03n (Exod. 12.11),
ownn wemisem ‘and equipped’ || MT owam (Exod. 13.18), onwn
ma'sim ‘smeared (with oil), anointed’ || MT nnwm (Lev. 2.4; 7.12; Num.
3.3; 6.15; 7.10), 531 négaf ‘struck (by plague), defeated’ || MT 4,3 ‘plague’
(Num. 8.19; Deut. 28.7, 25), Jwin anniisak ‘who is bitten’ || MT 73wan
(Num. 21.8; on the u-vowel in Samaritan Hebrew, see Ben-Hayyim
2000, 200, §2.3.14 Note), o°1n1 nétinam ‘given, dedicated’ || MT nin:
(Num. 8.16, 16, 19; 18.6; Deut. 28.31, 32), o-tpan affeqidam ‘those put
in charge, the officers’ || MT "tip (Num. 31.14, 48), ‘awn assebi ‘that
was captured’ || MT awn ‘the captivity’ (Num. 31.26), naw wsddifot
‘blighted (by the east wind)’ || MT ndi1w3 (Gen. 41.6, 23, 27). This com-
paratively high qatil/qatil incidence may well be due to the influence of
Samaritan Aramaic and of contemporaneous Hebrew dialects, e.g., RH,
in which gatil and qdtil served as passives with more regularity than in
Tiberian BH. The SP appearance of gétal passive forms of 175 and qn1 is
certainly a striking point of commonality with post-exilic Hebrew.

Tiberian ketiv-qere and the Samaritan Tradition 139

With this morphosemantic background in mind, it is possi-

ble to turn to the instance of ketiv-gere in (13).

(13) ny (Q) WM (K) WI pia Tan peng aT AWA Nw

DWW ANT

‘And Moses sent to spy out Jazer, and they captured its vil-

lages, so he dispossessed the Amorites who were there.’
(Num. 21.32)

There are opposing tendencies at work in the broader context of
this passage. On the one hand, excluding the verse under discus-
sion, throughout the book of Numbers, the seven cases of gal vw
(Num. 13.30; 21.24, 35; 27.11; 33.53; 36.8, 8) and the seven
cases of hifil win (Num. 14.12, 24; 32.21, 39; 33.52, 53, 55)
occur with their expected semantic and grammatical characteris-
tics, as described above, with the exception of the hifil form at
Num. 14.24. These include the nearly parallel usage to Num.
14.24 in the hifil with animate object Aa TW TAXATNY WI ‘and
he dispossessed the Amorite who dwelt therein’ (Num. 32.29).
On the other hand, in the immediate context of the ketiv-
qere dissonance of Num. 21.32—where, with an animate object,
one expects a hifil—come two cases of gal wy with inanimate
objects: izqx-nx wi) ‘and (Israel) took possession of his land’
(Num. 21.24) and igax-nsx iw ‘and they took possession of his
land’ (Num. 21.35). It would seem that the Tiberian ketiv reflects
a tradition in which the form of w">’ at Num. 21.32 was read as
qal in harmony with verbs in close proximity—resulting in a less
standard, but acceptable use of the qal with an animate object—

whereas the gere preserves a tradition more strictly observant of

140 Hornkohl

the standard semantic and grammatical distinction between qal
wy and hifil win.

Turning to the Samaritan tradition, one finds nearly the
same distribution of forms as in the Tiberian Torah, the chief dif-
ference being a greater number of gal forms due mainly to Samar-
itan textual pluses. Additionally, there are two individual cases of
what might be considered typical Samaritan harmonisation. First,
the lone Tiberian pi“el form in ‘All your trees and the fruit of your
land the 5y>xn (‘locust, cricket’?) will possess (wy") (Deut. 28.42)
is paralleled in the SP by hifil wiv juras. The ostensible hifil re-
placement of piel is certainly in line with the SP’s penchant for
levelling irregular forms, though hif‘l is unexpected in the case of
an inanimate object (though, to be sure, the non-human subject is
also exceptional).16 Second, the irregular Tiberian hifil with inan-
imate object suffix nwt “(his seed) will inherit it (i.e., the land)’
(Num. 14.24) finds as its Samaritan parallel the more predictable
gal with inanimate object suffix nw” yirdsinnd.

With the broader Samaritan picture and these individual ex-
amples in mind, it is no surprise that in the case of the Tiberian
ketiv-gere at Num. 21.32, the SP agrees with the tradition pre-

served in the Tiberian gere wi, albeit with slight adjustment for

'© Ben-Hayyim (1977, 130) analyses the form as a hif“il prefix conjuga-
tion, but it is equally analysable as a qal active participle, which would
preserve more conventional semantics and grammar. Incidentally, the
Tiberian piel may also be queried. As a hapax, one wonders if piel w3”
developed secondarily in place of qal w1” due to the unique usage. Al-
ternatively, the agricultural devastation wrought by an insect plague (if
that is what is envisioned) suits the ‘intensive’ semantics often associ-
ated with the piel stem.

Tiberian ketiv-qere and the Samaritan Tradition 141

purposes of number agreement with the closest subject referent,
i.e., WM wyurisu ‘and they dispossessed’; cf. the text and transla-
tion in example (13), above. The Samaritan imposition of order

and harmoniousness is thus conspicuous in several relevant facets.

8.0. Qere “n>? || ketiv 7) (Num. 23.13)

In the span of a few years, Steven Fassberg (1994, 13-35; 1999)
and Ahouva Shulman (1996, 65-84, see especially 84, n. 22) in-
dependently arrived at similar explanations for BH’s so-called
‘lengthened imperative’: that it denotes action in the direction of
the speaker or for the benefit thereof. Its use with such impera-
tives is not obligatory (see below), but is limited to such, at least
in most forms of ancient Hebrew, up to and including Second
Temple traditions, though there are sporadic signs of misuse and
definite signs of disuse (especially RH).'”

With specific regard to imperatival forms of gal 727, Shul-
man (1996, 75-81) argues convincingly that the short and long
forms normally denote, respectively, ‘go (away)’ and ‘come
(here)’, with the speaker as reference point. Shulman notes that
speaker-orientation is inferable from a following preposition with
first-person suffix and/or verb form, e.g., inclusive first-person plu-
ral cohortatives, first-person singular cohortatives denoting action
that can be performed only after the approach of the addressee,
and imperatives inviting action on behalf of the speaker. It must

be emphasised, though, that while long imperatives consistently

7 Arguable examples of archaising pseudo-classical misuse may be de-
tected in non-biblical material from Qumran, e.g., 4Q88 10.7, 8, 8;
4Q200 f5.9 (= Tobit 4.9); 4Q416 f£4.3; 4Q418 f222.2.

142 Hornkohl

denote speaker-orientation, the morphological marking is not ob-
ligatory for this speaker-orientation, e.g., the short form 79 is
sometimes followed by prepositions with first-person morphology
(e.g., Jdg. 18.19; 1 Kgs 13.15) or by first-person verbs implying
speaker-orientation (e.g., Jdg. 4.22; 2 Chron. 25.17 ketiv).

We are now positioned to examine the ketiv-gere instance

in example (14).

ree

‘And Balak said to him, “Please come with me to another
place, from which you may see them. You shall see only a
fraction of them and shall not see them all. Then curse them
for me from there.”’ (Num. 23.13)

That the imperative 75 invites movement in the direction of the
speaker is indicated by the following *mx ‘with me’ and *9-i1271
own ‘and curse him for me from there’. Given the examples of
speaker-oriented short-imperative 74, above, the ketiv must be
seen as an acceptable, if rare, use of the short imperative for ex-
pressing movement toward the speaker. The gere, conversely, re-
flects the more common lengthened morphology of the impera-
tive, -n29, in the sense of ‘come’ (cf. the ketiv-gere in 2 Chron.
25.17). The gere is also in line with parallel commands in the
context (Num. 10.29; 22.6, 11, 17; 23.7, 27; 24.14; cf. the short
imperative in cases of ablative movement: Num. 22.20, 35).

The Samaritan situation is complex. On the one hand,
where it appears in the SP, use of the lengthened imperative re-

sembles that in the MT. This is to say that the lengthened imper-

Tiberian ketiv-qere and the Samaritan Tradition 143

ative appears in Samaritan Hebrew in the same grammatical con-
texts and with the same meaning as in Tiberian Hebrew, i.e., for
actions involving motion toward the speaker or for the benefit
thereof. Even so, the lengthened imperative is less common in the
SP than in the MT Torah. This tallies with the aforementioned
disuse of the form in some forms of late antique Hebrew, most
notably RH. Indeed, against just four cases in which the SP has a
lengthened imperative and the MT does not (Gen. 19.9; Exod.
17.2; Num. 21.6; Num. 23.18), there are twenty or more arguable
cases in which the SP has a short form against an MT lengthened
one (Gen. 15.9; 19.32; 21.23; 25.33; 27.3, 4, 7; 29.19, 21;'°
37.13; 43.8; 47.31; Exod. 3.10; 32.10;!9 Num. 10.29; 23.13, 27;
24.14; 27.4; Deut. 26.15; 33.23). Significantly, seven of these lat-
ter involve Samaritan 7 lik || MT 729. Though the combined Sa-
maritan written-reading tradition preserves lengthened impera-
tives, in general, and the lengthened 755 lika, more specifically
(Gen. 31.44; Num. 22.6, 11, 17; 23.7, 7), the SP seems to evince
a situation in which the perceived distinction between short and
lengthened imperatives has undergone a degree of erosion, so
that retention of the final n- -a was not deemed vital for the sake

of semantic disambiguation.”

’® Here 737 is read ibi = Tiberian xan ‘bring’.

' The SP reads * Amin anniy'yé-lli (Ben-Hayyim 2000, 74, §1.4.10), but
it is not clear that the -é suffix is that of the lengthened imperative.

° This contrasts markedly with Samaritan use of the lengthened first-
person wayyiqtol, i.e., the pseudo-cohortative, which, in line with other
Second Temple Hebrew chronolects, is far more common in both the
written and reading components of the SP than in the Tiberian Penta-
teuch (see Hornkohl, f.c.).

144 Hornkohl

Parallel to the Tiberian ketiv-gere in Num. 23.13, the SP has
unlengthened 75 lik, in agreement with the Tiberian ketiv. This
concords with the same form at SP Num. 23.27 and 24.14 (both
of which have lengthened imperatives in the MT), but clashes
with lengthened 735 lika at SP Num. 22.6, 11, 17; 23.7, 7 (which
are all lengthened in the MT). Such diversity, especially in a sin-

gle pericope, is uncharacteristic for the SP.

9.0. Qere win || ketiv prin (Num. 32.7)

(15) WWy PISA oy Dau DN Iw 22 Tey (Q) PNET (K) PNTIN nN

smi 07? 13

‘Why will you discourage the heart of the people of Israel

from going over into the land that the LorD has given
them?’ (Num. 32.7)

Since all other forms of the verb in question in the Tiberian tra-
dition are hifil (Num. 30.6, 6, 9, 12; 32.9; Ps. 33.10; 141.5)—
including consonantally unambiguous qatal forms (Num. 30.6, 6,
12; Ps. 33.9) and a form in the immediate vicinity in the same
idiom 5xyw> 332 39-nx 18731 ‘and they discouraged the hearts of the
children of Israel’ (Num. 32.9)—it is difficult to view the ketiv in
example (15) as anything other than a result of conflation of waw
and yod, presumably arising from their graphic similarity.

The verb in the SP is consistently hifil, with no divergences

between the written and reading components of the tradition

Tiberian ketiv-qere and the Samaritan Tradition 145

(Num. 30.6,7! 6, 6, 12; 32.7, 9). Thus, in Num. 32.7, SP xin
tanniyyon || MT gere 18°15.

10.0. Qere 3cPL qatal 1- || ketiv 3FPL qatal n- (Deut.
21.7; Num. 34.4?)

In the lower margin of L on the page that includes Deut. 21.7 (fol.

111v), the masora parva reads:

ndy wan nya mawica naw xb Amp Amp xd fh Api A Aa 7
moan ADaW nAAW TY Ma xXdi qr Any naw xd Anyi py
ndbw artiy

Fourteen™ times the ketiv is heh and the gere is waw: -x9
nn? ANP ‘they shall not make bald patches’ (Lev. 21.5);
maw x2 ‘(our hands) did not spill’ (Deut. 21.7); Aaw3D
‘for (the ships) were wrecked’ (1 Kgs 22.49); a>y x70 nya
‘at that time (the servants of Nebuchadnezzar king of Bab-
ylon) came up’ (2 Kgs 24.10); nny3 wy ‘his cities fell to
ruin’ (Jer. 2.15); nawii x9 ‘(cities that are) not settled’ (Jer.
22.6); m1r nAY ‘now they will prostitute’? (Ezek. 23.43);
Tiy-marrxd) ‘and they will no longer be’ (Ezek. 37.22);
nanw ‘they have been laid desolate’ (Ezek. 35.12); naaw
‘(my feet) slipped’ (Ps. 73.2); nan7nn ‘(my face) reddened’
(Job. 16.16); nny ‘(our eyes) still’ (Lam. 4.17); n>w ‘blas-
phemy’ (Dan. 3.29)

21 SP Num.30.6 includes an infinitive absolute x17 anni with no parallel
in the MT.

2 The note gives the figure 7° ‘fourteen’, but lists just thirteen cases,
omitting all five occurrences of mixin 7°71 (see below) as well as {xv
‘AY m7 Hitax ‘lost sheep have been my people’ (Jer. 50.6).

*° The ketiv in this verse is actually mir ny, the gere 1r ANY.

146 Hornkohl

While the forms listed in the masoretic note represent various
categories with diverse explanations for the interchange, several
involve an apparent 3Fs qatal form and a FPL or FDL subject. There
is consensus that the Proto-Semitic suffix conjugation paradigm
distinguished between 3MPL and 3FPL endings, the former -u and
the latter -@ (Huehnergard and Pat-El 2019, 8). A distinction is
observed in Akkadian (-i vs. -@; Hasselbach-Andee 2019, 105),
Arabic (-a vs. -na; Birnstiel 2019, 384), Aramaic (-u vs. -u/a/in;
Kaufman 1998, 126), Syriac ([-w] vs. [-y]; Pat-El 2019, 663), and
Ge‘ez (-u vs. -a; Butts 2019, 132). By contrast, in Hebrew, for the
most part, dedicated 3FPL qatal morphology fell out of use in fa-
vour of epicene 1-. There are some 25 cases in which it has been
argued that biblical suffix conjugation forms ending in 7%:- or
ketiv n- with plural subjects represent a form with dedicated 3FPL
morphology (see Hornkohl 2013, 142-45 for summary and ref-
erences). This seems a plausible explanation for the ketiv form in

example (16) and may also have relevance for (17).

(16) xq x8>arry) ain oFA-nex (Q) Day (K) TEU x9 arp Nt 1
‘and they shall testify, ‘Our hands did not shed this blood,
nor did our eyes see it shed.’ (Deut. 21.7)

In L, intercolumnal notes on ir7? at Deut. 21.7 and iry? at Jer.
6.24 read 5n j ‘six times plene’.*4 This shows, among other things,

that the Masoretic Tradition was primarily concerned with the

74 According to L, there are actually seven such instances: Gen. 5.29;
Deut. 21.7; 32.27; Jer. 6.24; Hos. 14.4; Ps. 90.17, 17. However, the
masoretic note is confirmed by similar notes in A at Hos. 14.4 and Ps.
90.17 and, most crucially, by A’s defective 1177 at Deut. 32.27, which
shows that L deviates slightly from its own masora.

Tiberian ketiv-qere and the Samaritan Tradition 147

word’s correct orthography, whatever its meaning. In this case,
the ketiv na5w can be argued to preserve an archaic 3FPL in agree-
ment with FDL ‘our hands’, while the gere unambiguously repre-
sents the more standard 3cPL, effecting harmony with the 3cPL in
1x7 XD ippyi ‘and our eyes did not see’ later in the verse.

The SP not surprisingly adopts the more standard cDL x3 177
yaw yedinu 1d sdfdku ‘our hands did not shed’. This is in keeping
with the tradition’s quest for consistency. Yet, the Samaritan pro-
pensity for levelling unconventional forms does not preclude the
possibility of preserved archaisms. Indeed, MT 1123 5x7" "ry ‘and
Israel’s eyes grew heavy (= dim)’ (Gen. 48.10) || SP at29 Saw ryn
wini yisrd’al kabdda, where the SP verb is evidently a suffix conjuga-
tion (gatal) form preserving dedicated 3FP morphology.

The above considerations may also apply to example (17).

(17) (Q) P71 (K) TT yg ray brarpy nova? agin ddan be? apn
*AYOLY TAY] TIN WT kN INA Wp? WIN PORN

‘And your border shall turn south of the ascent of Akrab-

bim, and cross to Zin, and its limit shall be south of
Kadesh-barnea. Then it shall go on to Hazar-addar, and pass

along to Azmon.’ (Num. 34.4)

In BH, the lexeme nixyin ‘limits, farthest reaches’ (25x) is always
plural. In (17) it is tempting to attribute the apparent mismatch
between ketiv 7m and plural subject rnxyin ‘its limits’ to an al-
ternative syntactic interpretation, according to which 7°m func-
tions as a ‘discourse marker’ rather than a verb proper (see van
der Merwe, Naudé, and Kroeze 2017, 427-28). In that case, a
corresponding English translation would have a discourse mark-

ing ‘and it will be’ followed by the rendering of a verbless clause

148 Hornkohl

‘its limits will be south of....’ Yet, it is important to consider this
case from a broader perspective. A form of nixyin follows a form
of 7° nineteen times in BH. In eleven of these 77 is plural (Num.
34.5, 8, 9, 12; Josh. 15.7, 11; 16.3, 8; 19.14, 22, 29), in three 77
is singular (Josh. 17.9, 18; 19.33), and in five the verb is singular
in the ketiv and a plural in the gere (Num. 34.4; Josh. 15.4; 18.12,
14, 19). This means that according to the written component of
Tiberian BH, the form nixyin is the subject of an apparent singu-
lar form of 7° nearly as often (8x) as it is the subject of a plural
form of the verb (11x). In six of the eight cases of apparently
singular 77, the verb is a suffix conjugation form; in the other
two, the verb is *7 (Josh. 17.9; 19.33). It is worth mentioning at
this point that though the Tiberian tradition regularly construes
nxyxin as a plural via verbal agreement and/or a plural possessive
suffix 1- (cf. ketiv inxyn Josh. 16.3), only in a minority of cases is
the form explicitly spelled as a plural in n\-. It may be that, along-
side the plural form, a singular along the lines of AX¥Nn* or nR¥n*
was also known (see below), but was secondarily levelled in con-
formity with the plural at a date sufficiently early that the plu-
rality was sporadically recorded in the spelling tradition.
Whatever the exact explanation for the ketiv-gere disso-
nance in (17) and the other four relevant ketiv-gere instances in-
volving 77 and mixyn, it is clear that the written component of
the Tiberian tradition preserves a situation of singular and plural
diversity more extensive than that preserved in the correspond-
ing reading tradition, where plural agreement is greatly, albeit

not exclusively, favoured.

Tiberian ketiv-qere and the Samaritan Tradition 149

Intriguingly, the SP agrees with the Tiberian ketiv: mm
wéyya. This is in line with the Samaritan convention according to
which the noun imxxin tisd@itu is treated as singular and consist-
ently paired with 3ms nm. This may be considered evidence for
an originally singular option for consistently plural Tiberian
nixyn (see above).

11.0. Qere r2w || ketiv 5xw (Deut. 28.30)

The euphemistic employment of 22 ‘lie (down)’ in reference to
sexual relations is common throughout BH (and is matched by
euphemistic renderings in the ancient versions). This usage was
also secondarily extended to instances of ketiv 5"3w ‘rape, ravage’,
one such case obtaining in example (18) (see also Isa. 13.16; Jer.
3.2; Zech. 14.2).

(18) awn-xd) nian ma (Q) map" (K) mb3w* Sng WN] WINN AWN
‘You shall betroth a wife, but another man shall ravish
her. You shall build a house, but you shall not dwell in it.
You shall plant a vineyard, but you shall not enjoy its fruit.’
(Deut. 28.30)

Cohen (2007, 264) proffers a compelling motivation for such eu-
phemistic ketiv-gere cases. Words that were deemed problematic
to utter in public, due to perceived impropriety or taboo, were
replaced in oral recitation by more appropriate substitutes, but
continued to be copied faithfully in the written tradition.

In the case of (18) and similar, the euphemistic substitution

could not be effected without certain grammatical modifications.

150 Hornkohl

First, the verb 12 normally takes one of the comitative preposi-
tions, OY or nx both ‘with’ (Orlinsky 1944). On seven occasions
one encounters -nx 13¥, i.e., the definite accusative/direct object
marker, but in six of the seven only the vocalisation calls for such
an analysis.*° The earlier syntax was more likely with the prepo-
sition nx ‘with’, its reinterpretation as the direct object marker
secondary. In this way, secondary disambiguation was created
between originally intransitive 13 with comitative Dv or nx ‘lie
with’, on the one hand, and the innovative transitive 122¥ with
accusative nx in the more aggressive sense of ‘forcibly engage in
sex’, on the other (cf. the Targumic distinction between o’y 19
and mn’ 15w). Relatedly, the verb 19W nowhere in BH bears an ob-
ject suffix except where it is read as the gere for ketiv 53w, as in
(18) above. Finally, BH lacks a nif‘al 12W1 except where it is read
instead of apparently nif‘al 51w3*, as in Isa. 13.16 and Zech. 14.2.
Significantly, unambiguous consonantal nif‘al 13w1" is first at-
tested in material in the non-biblical material from Qumran
(4Q270 f5.19; 4Q271 £3.12) and persists in RH. Relatedly, no
passive qal or pu“al cognate of 13 is known from ancient He-
brew beyond that in the gere of Jer. 3.2 (and no piel is attested
at all). All of the above point to the likely secondary development
of transitive -nx 12, perhaps in the early Second Temple period
(cf. -nix 12v with mater waw in Ezekiel) (Beuken 2004, 663). In
other words, -nx 13W is itself an unmarked case of ketiv-qere mis-
match in line with the replacement of transitive 5"3w with origi-

nally intransitive 1"2w for purposes of (public) oral recitation.

25 -nx: Gen. 34.2; Lev. 15.18, 24; Num. 5.13, 19; 2 Sam. 13.14; -nix:
Ezek. 23.8.

Tiberian ketiv-qere and the Samaritan Tradition 151

The gere form in (18) involves two of the three aforementioned
secondary developments: 13w with transitive semantics and 13
with an object suffix—both traits that it seems to have inherited
due to its substitution for transitive (presumably G-stem) 5"3w.

While gere 1">w is almost certainly secondary, the evidence
seems indicative of rather early replacement. Greek é£e adrjv
‘will have her’ and Syriac guns ‘will take her’ are ambiguous as
evidence of their Hebrew Vorlage. Though certainly euphemistic,
they do not obviously correspond to either the Tiberian ketiv or
qere. But other ancient versions arguably confirm the antiquity of
the gere tradition: Vulgate dormiat cum ea ‘will sleep with her’;
Targum Onkelos nra5v” ‘will lie with her’; 1QIsa* 11.24 mAiaSwn
|| MT Isa. 13.16. Depending on the antiquity of its plene spelling,
the Tiberian orthographic -mix 19w in Ezek. 23.8 may also testify
to the antiquity of the substitution.

For its part, the combined Samaritan written-reading tradi-
tion at Deut. 28.30 has nny anv yiskab imma ‘will lie with her’.
Assuming the primary status of 7351”, as in the Tiberian ketiv,
the Samaritan euphemistic solution goes farther than that of the
Tiberian gere. It avoids not only an inappropriate word, but tor-
tured grammar, too, resorting to conventional rection for the
verb 12w with a transparently comitative preposition. Indeed, Sa-
maritan Hebrew does not know the formulation of transitive 1.w
with direction object nx; Tiberian -nx 13w is consistently paral-
leled by Samaritan -nx 12w Sdkdb itt- (Gen. 34.2; Lev. 15.18, 24;
Num. 5.13, 19), i.e., intransitive verb with comitative preposi-
tion. In this case the Samaritan penchant for harmonisation has

led to mixed results of development and conservation: on the one

152 Hornkohl

hand, modification of the original 7151” to more acceptable and
grammatical AAy 13v” not only in the recitation tradition, but at
the level of the orthography; on the other hand, preservation of
aw with a comitative preposition, in contrast to the innovative
Tiberian distinction between neutral 13w with comitative -nx ‘lie
with’ and the more explicitly non-consensual 13W with accusa-

tive/direct object -nx ‘rape, ravish’.

12.0. Conclusion

In the introduction to this study, the diachronic relationship be-
tween the various relevant linguistic traditions of the Torah were
sketched as follows: an ancient Tiberian orthographic compo-
nent; a largely harmonious, but somewhat later Tiberian reading
component; an ancient Samaritan written component with clear
and widespread evidence of Second Temple reworking; and a Sa-
maritan reading component replete with Second Temple and
later features. While this broad characterisation may be generally
accurate, it finds only partial support in the cases examined
above; see Table 3. This fact should inform understanding of the
relationship between the Tiberian ketiv and gere traditions.

A few of the cases discussed in the body of this study ex-
hibit the diachronic progression expected based on the charac-
terisation sketched in the introduction, e.g., archaic Tiberian
ketiv, standard Tiberian gere, and late Samaritan combined tradi-
tion: 8810.0 (though the SP elsewhere also preserves such an ar-
chaism) and 11.0. Similar is the situation of archaic ketiv, late
qgere, and later SP in §2.0 as well as those of the unexplained ketiv,

standard gqere, and late SP form in 83.0 and the SP’s levelling in

Tiberian ketiv-qere and the Samaritan Tradition 153

favour of a late form in §6.0. The relatively late character of the

SP is the most conspicuous diachronic trait, though its preserva-

tion of gal morphology against Tiberian gere nif‘al in the case of

§5.0 might be an exception, as may more than one Samaritan

feature associated with §10.0.

Table 3: Ketiv, gere, and Samaritan findings with summary discussions

1.0.

Q: RYT K: xxin SP: x7xin tsi
qgere: non-standard I-y (error? euphony?), ketiv: standard I-w; SP = ketiv

2.0.

Q: own K: ow SP: ow" wyuwwdsdm
qere: late, ketiv: archaic; SP: late Aramaic/post-biblical stem

3.0.

Q: ana" K: innwn SP: nnn» yistdbbu
qere: standard plural, ketiv: unexplained (?); SP = gere, with later phonology

4.0.

Q: Wa K: 731 SP: 731 afgad

qere/ketiv: plausible; SP = ketiv, ST = qere

5.0.

Qrartey; abn; arb K: wos; indn; wd SP: 1d" wyilldnu; wn tillanu; 15" wyillanu
qere/ketiv: complex stem arrangement; SP = qere/ketiv; less differenti-
ated arrangement; shared gemination and qal vestiges possibly ancient

6.0.

QoNTB RTP Kor py xmp SP: xp garya@’i
qere and ketiv agree on variation, but not location; SP alternately =
ketiv/gere, unifying according to a late pattern typical of Aramaic/RH

7.0.

Q: wri" Ki vv SP: wm wyurisu
gere: global morphosemantic consistency; ketiv local harmony; SP =
qere, with broad morphosemantic consistency

8.0.

Q: “725 K: > SP: 75 lik
qere: standard marked usage, ketiv: acceptable unmarked variant; SP =
ketiv, along with less use of the marked option and local inconsistency

9.0.

Q: pein K: xin SP: yiNkIn tanniyyon
qere: standard, ketiv: graphic error (?); SP = qere

10.0.

Q:isaw; qm «Ki naaw;arm SP: paw sdfdku; mm weyya
qere: standard 3cpL ending, ketiv: archaic 3FPL ending (frequent with
mixin); SP = gere/ketiv; SP knows the archaic 3FPL

11.0.

Q: maw K: now SP: nay raw yiskab imma
qere: late euphemistic replacement, with syntax of ketiv, creating distinct
sense of 19W; SP = qere, with syntax of substitute lexeme

Beyond this, it is worth remarking that the SP agrees with

the Tiberian ketiv nearly as often—five occasions: §81.0, 4.0

(against the ST), 8.0, 10.0—as it agrees with the Tiberian gere—

154 Hornkohl

six occasions: §§3.0, 6.0, 7.0 (with slight modification), 9.0, 10.0,
11.0. This seems due mainly to the Samaritan penchant for con-
sistency and harmony, which often leads to levelling in line with
the majority form, which is the ketiv in 881.0 and 10.0 (7m), but
the gere in §§3.0, 7.0, 9.0, 10.0 (325w), and 11.0. Be that as it
may, the SP occasionally exhibits inconsistency and/or a minor-
ity form: 88.0. Moreover, the non-uniform character of the Sa-
maritan exegetical tradition is evidenced by the divergent inter-
pretations of the SP and the ST in 84.0.

Turning to the combined Tiberian tradition, while inexpli-
cable forms are occasionally presented by both the ketiv and the
qere, it seems that in the majority of cases the preserved form in
each tradition can be justified. More rarely—especially in in-
stances where graphic similarity between waw and yod may have
been at play—it seems likely that the ketiv form represents a cor-
ruption avoided in the gere (883.0, 9.0)—though, it has been sug-
gested that an otherwise unexplained gere form may have arisen
from waw-yod conflation: §1.0.

It has been remarked that in most cases the Tiberian ketiv
and gere both represent plausible readings. While this arguably
sheds important light on the ketiv-gere phenomenon, there is ev-
idence that the traditions differ with respect to more than just
natural historical linguistic development. Consider, in particular,
the cases discussed in §87.0 and 11.0. In both cases, the qere
seems to reflect a linguistic tradition characterised by deliberate
care. In §7.0, this manifests in the global morphosemantic con-

sistency of the distinction between nearly synonymous hif‘“l and

Tiberian ketiv-qere and the Samaritan Tradition 155

qal forms. In 811.0, it is seen in secondary disambiguation be-
tween intransitive (comitative) 13 and transitive 12W. Both de-
velopments reflect what in another connection Khan (2021a,
330-31) has described as “a general Second Temple development
in the proto-Masoretic reading tradition involving the introduc-

tion of strategies to increase... clarity of interpretation.”
References

Primary Sources

A= Aleppo Codex: Bar Hama, Adon. 2002. Digitalization of The
Aleppo Codex. https://barhama.com

Ben Sira: Abegg, Martin G. 2009. Ben Sira Electronic Database
(BENSIRA-C/BENSIRA-M). Accordance Modules.
Altamonte Springs, FL: OakTree Software, Inc.

Dead Sea Scrolls Biblical Texts: Abegg, Martin G., Jr. 2009a. Dead
Sea Scrolls Bible Canonical/Manuscript Order (DSSB-C/DSSB-
M). Accordance Modules. Altamonte Springs, FL: OakTree
Software, Inc.

Dead Sea Scrolls Non-biblical Texts: Abegg, Martin G., Jr. 2001.
Non-Biblical Dead Sea Scrolls from Qumran (QUMRAN). Ac-
cordance Module. Altamonte Springs, FL: OakTree Soft-
ware, Inc.

Egyptian Aramaic: Porten, Bezalel, and Ada Yardeni (eds.). 2007.
Textbook of Aramaic Documents from Ancient Egypt (TAD-T).

6 Khan subsumes such strategies under the general heading of orthoepy;
see Khan (2018b; 2020, I: 73-85, 99-105).

156 Hornkohl

Accordance Module. Altamonte Springs, FL: OakTree Soft-
ware, Inc.

L = Leningrad Codex: Zuckerman, Bruce E. 1990. Leningrad Co-
dex Images. Included in BibleWorks 10. West Semitic Re-
search Group.

LXX: Taylor, Bernard A., and Dale M. Wheeler (eds.). 2008. Greek
Septuagint Database. Accordance Module. Altamonte
Springs, FL: OakTree Software, Inc. Based on Ralfs, Alfred
(ed.). 2006. Septuaginta. Revised by Robert Hanhart.
Stuttgart: Deutsche Bibelgesellschaft.

Peshitta: Abegg, Martin, and Jerome Lund (eds.). 2010. Syriac
Peshitta (Old Testament) (PESHOT-T). Accordance Module.
Altamonte Springs, FL: OakTree Software, Inc.

RH = Rabbinic Hebrew (Mishna, Tosefta, Talmuds): Ma’agarim
Database of the Academy of the Hebrew Language:
https://maagarim.hebrew-academy.org.il/Pages/PMain.aspx

Samaritan Arabic Translation: Shehadeh, Haseeb (ed.). 1989-
2002. The Arabic Translation of the Samaritan Pentateuch:
Edited from the Manuscripts with an Introductory Volume. 2
vols. Jerusalem: Israel Academy of Sciences and Humani-
ties. [Arabic and Hebrew]

Samaritan Pentateuch Oral Tradition: Ben-Hayyim, Ze’ev. 1977.
The Words of the Pentateuch. Vol. 4 of The Literary and Oral
Tradition of Hebrew and Aramaic amongst the Samaritans. Je-

rusalem: The Academy of the Hebrew Language. [Hebrew]

Tiberian ketiv-qere and the Samaritan Tradition 157

Samaritan Pentateuch Written Tradition: Tal, Abraham, and
Moshe Florentin (eds.). 2010. The Pentateuch: The Samari-
tan Version and the Masoretic Version. Tel-Aviv: The Haim
Rubin Tel-Aviv University Press.

Samaritan Targum: Tal, Abraham. 2010. Samaritan Targum
(SAMTARG). Accordance Module. Altamonte Springs, FL:
OakTree Software, Inc.

Targums: Cook, Edward M. (ed.). 2007a. Targum Geniza (TARGG-
T). Accordance Module. Altamonte Springs, FL: OakTree
Software, Inc.

. 2007b. Targum Fragments (TARGF-T). Accordance Mod-

ule. Altamonte Springs, FL: OakTree Software, Inc.

2008. Targum (TARG-T). Accordance Modules.
Altamonte Springs, FL: OakTree Software, Inc.

Vulgate: Fabbri, Marco V. 2008. Tagged Latin Vulgate (VULG-T).
Accordance Modules. Altamonte Springs, FL: OakTree Soft-
ware, Inc. Based on Weber, Robert (ed.). 1994. Biblia Sacra
Tuxta Vulgatem. 4th edition. Stuttgart: Deutsche Bibelgesell-
schaft.

Secondary Sources

Ben-Hayyim, Ze’ev. 2000. A Grammar of Samaritan Hebrew: Based
on the Recitation of the Law in Comparison with the Tiberian
and Other Jewish Traditions. Revised English edition. Jeru-
salem: The Hebrew University Magnes Press.

Bergstrasser, Gotthelf. 1918-1929. Hebrdische Grammatik. 2 vols.
Leipzig: Vogel.

158 Hornkohl

Beuken, Willem A. M. 2004. ‘13’. In Theological Dictionary of the
Old Testament, edited by G. Johannes Botterweck, Helmer
Ringgren, and Heinz-Josef Fabry, translated by Douglas W.
Scott, vol. 14, 659-71. Grand Rapids, MI: Eerdmans.

Birnstiel, Daniel. 2019. ‘Classical Arabic’. In The Semitic Lan-
guages, 2nd edition, edited by John Huehnergard and
Na‘ama Pat-El, 367-402. London: Routledge Taylor &
Frances Group.

Blau, Joshua. 2010. Phonology and Morphology of Biblical Hebrew.
Linguistic Studies in Ancient West Semitic 2. Winona Lake,
IN: Eisenbrauns.

Butts, Aaron Michael. 2019. ‘GaYaz (Classical Ethiopic)’. In The
Semitic Languages, 2nd edition, edited by John Huehnergard
and Na‘ama Pat-El, 117-44. London: Routledge Taylor &
Frances Group.

Cohen, Maimon. 2007. The Kethib and Qeri System in the Biblical
Text: A Linguistic Analysis of the Various Traditions Based on
the Manuscript ‘Keter Aram Tsova’. Jerusalem: The Hebrew
University Magnes Press. [Hebrew]

Fassberg, Steven E. 1994. Studies in Biblical Syntax. Jerusalem:
The Hebrew University Magnes Press. [Hebrew]

. 1999. ‘The Lengthened Imperative n9vp in Biblical He-

brew’. Hebrew Studies 40: 7-13.

Fox, Joshua. 2003. Semitic Noun Patterns. Harvard Semitic Studies

52. Winona Lake, IN: Eisenbrauns.
GKC = Kautzch, Emil (ed.). 1910. Gesenius’ Hebrew Grammar.
Translated by Arthur E. Cowley. Oxford: Oxford University

Press.

Tiberian ketiv-qere and the Samaritan Tradition 159

Hasselbach-Andee, Rebecca. 2019. ‘Akkadian’. In The Semitic Lan-
guages, 2nd edition, edited by John Huehnergard and
Na‘ama Pat-El, 95-116. London: Routledge Taylor &
Frances Group.

Hornkohl, Aaron D. 2013. Ancient Hebrew Periodization and the
Language of the Book of Jeremiah. Studies in Semitic Lan-

guages and Linguistics 74. Leiden: Brill.

. 2018. ‘Diachronic Exceptions in the Comparison of Tibe-
rian and Qumran Hebrew: The Preservation of Early Linguis-
tic Features in Dead Sea Scrolls Biblical Hebrew’. In The Re-
configuration of Hebrew in the Hellenistic Period: Proceedings of
the Seventh International Symposium on the Hebrew of the Dead
Sea Scrolls and Ben Sira at Strasbourg University, June 2014,
edited by Jan Joosten, Daniel Machiela, and Jean-Sébastien
Rey, 61-92. Studies on the Texts of the Desert of Judah 124.

Leiden: Brill. https://doi.org/10.1163/9789004366770_006
. 2020a. ‘Discord between the Tiberian Written and Reading

Traditions: Two Case Studies’. In Studies in Semitic Vocalisa-
tion, edited by Aaron D. Hornkohl and Geoffrey Khan, 227-
80. Cambridge: University of Cambridge and Open Book Pub-
lishers. https://doi.org/10.11647/OBP.0207.07

. 2020b. ‘The Hebrew of the Dead Sea Scrolls and the Ti-
berian Reading Tradition: Shared Departures from the Mas-
oretic Written Tradition’. Dead Sea Discoveries 27/3: 410-

25. https://doi.org/10.1163/15685179-bja10012
. 2021. ‘Niphalisation in Ancient Hebrew: A Perspective

from the Samaritan Tradition’. Journal for Semitics 30/2: 1-
17. https://doi.org/10.25159/2663-6573/9207

160 Hornkohl

. Forthcoming. Studies in the Historical Depth of the Tiberian
Reading Tradition. Cambridge Semitic Languages and Cul-
tures. Cambridge: The Faculty of Asian and Middle Eastern
Studies and Open Book Publishers.

Huehnergard, John. 2007. ‘Qatil and Qatil Nouns in Biblical He-
brew’. In Sha‘arei Lashon: Studies in Hebrew, Aramaic, and
Jewish Languages Presented to Moshe Bar-Asher, edited by
Aharon Maman, Steven E. Fassberg, and Yochanan Breuer,
I:*3-*45. Jerusalem: Bialik Institute.

Huehnergard, John, and Na‘ama Pat-El. 2019. ‘Introduction to
the Semitic Languages and Their History.’ In The Semitic
Languages, 2nd edition, edited by John Huehnergard and
Na‘ama Pat-El, 1-21. London: Routledge Taylor & Frances
Group.

Kaufman, Stephen A. 1998. ‘Aramaic’. In The Semitic Languages,
edited by Robert Hetzron, 114-30. London: Routledge Tay-
lor & Frances Group.

Khan, Geoffrey. 201 3a. ‘Ketiv and Qere’. In Encyclopedia of Hebrew
Language and Linguistics, edited by Geoffrey Khan, II: 463-
68. Leiden: Brill.

. 2013b. ‘Tiberian Reading Tradition’. In Encyclopedia of

Hebrew Language and Linguistics, edited by Geoffrey Khan,
III: 769-78. Leiden: Brill.

.2018. ‘Orthoepy in the Tiberian Reading Tradition of the
Hebrew Bible and Its Historical Roots in the Second Temple
Period’. Vetus Testamentum 68: 1-24.

. 2020. The Tiberian Pronunciation Tradition of Biblical He-

brew. 2 vols. Cambridge Semitic Languages and Cultures 1.

Tiberian ketiv-qere and the Samaritan Tradition 161

Cambridge: University of Cambridge and Open Book Pub-

lishers. https://doi.org/10.11647/OBP.0194.01
. 2021. ‘The Coding of Discourse Dependency in Biblical He-

brew Consecutive Weqatal and Wayyiqtol’. In New Perspectives
in Biblical and Rabbinic Hebrew, edited by Aaron D. Hornkohl
and Geoffrey Khan, 299-354. Cambridge Semitic Languages
and Cultures 7. Cambridge: University of Cambridge and
Open Book Publishers. https://doi.org/10.11647/OBP.0250

Morgangw, Lolo. 1858. ‘History of the British Bards: Rhyme, An-
tiquity of. The Cambrian Journal, 353-63. Published under
the auspices of the Cambrian Institute, London: Longmans
& Co., J. Russel Smith, and J. Petheram.

Orlinsky, Harry M. 1944. ‘The Hebrew Root SKB’. Journal of Bib-
lical Literature 63/1: 19-44.

Pat-El, Na‘ama. 2019. ‘Syriac’. In The Semitic Languages, 2nd edi-
tion, edited by John Huehnergard and Na‘ama Pat-El, 653-
78. London: Routledge Taylor & Frances Group.

Shulman, Ahouva. 1996. ‘The Use of Modal Verb Forms in Bibli-
cal Hebrew Prose’. PhD dissertation, University of Toronto.

Tov, Emanuel. 2012. Textual Criticism of the Hebrew Bible. 3rd re-
vised and expanded edition. Minneapolis, MN: Fortress
Press.

van der Merwe, Christo H. J., Jacobus Naudé, and Jan H. Kroeze.
2017. A Biblical Hebrew Reference Grammar. 2nd edition.
London: T&T Clark, Bloomsbury Publishing.

Yalon, Hanoch. 1971. Pirge Lashon. Jerusalem: Bialik.

162 Hornkohl

Yeivin, Israel. 1980. Introduction to the Tiberian Masorah. Trans-
lated and edited by E. J. Revell. Masoretic Studies 5. Mis-
soula, MT: Scholars Press, for The Society of Biblical Liter-
ature and the International Organization for Masoretic
Studies.

A FURTHER ANALYSIS OF THE
‘BYZANTINE (ITALIAN-LEVANTINE)
TRIAD’ OF FEATURES IN COMMON

TORAH CODICES’

Estara J Arrant

In my recent studies on the variation of Tiberian vowel and dia-
critic signs in medieval Hebrew Bible codices from the Cairo Ge-
nizah, I have highlighted, analysed, and contextualised a specific
pattern involving the Tiberian signs shewa and dagesh (Arrant
2020; 2021). This pattern of features, which in this article is

called the ‘Byzantine Triad’ of features,” includes the following:

e the placement of a sign resembling dagesh in consonantal
-alef, often with a corresponding rafe placed over mater lec-

tionis °alef;?

' Many thanks to the editors and peer reviewers of this volume for their
helpful comments. I thank the Syndics of the Cambridge University Li-
brary for permission to use the images of the manuscripts which appear
here.

? In previous studies (Arrant 2020; 2021), I called this phenomenon the
‘Byzantine Trio’.

Typically, this is accompanied by a pattern of rafe usage extended to
non-begedkefet letters, but this is not further discussed here, as variation

164 Arrant

e the placement of shewa under otherwise unvocalised word-
final ‘ayin and het;

e a pattern of ‘extended’ use of dagesh forte in letters which
do not, according to the standard rules of the Tiberian sys-

tem, require a dagesh forte.

These variations have been discussed previously in the context of
‘Palestino-Tiberian’ vocalisation and ‘extended Tiberian’ vocali-
sation and have been identified in famous codices (such as Codex
Reuchlinianus) (Diez Macho 1956; 1963; Morag 1959; Yeivin
1983; Fassberg 1990; Khan 1991; 2017; 2020; Heijmans 2013;
Blapp 2017). However, the discussions are somewhat limited in
focus to the developmental chronology of these particular sys-
tems of sign usage within the Tiberian Masoretic tradition, treat-
ing these features individually, rather than in conjunction with
each other.

Prior to Arrant (2020; 2021), the significance of the specific
pattern of co-occurrence of these three signs had gone unnoticed
in scholarship. In these two studies, I took a contextualising ap-
proach to the vocalisation of Geniza Torah codices and, through
the use of machine learning algorithms, analysed a large swath
of around 1,800 Torah codices with many different kinds of non-
standard Tiberian vocalisation. In a sub-group of the corpus, I
identified the three features listed above, which co-occur in a dis-
tinct pattern (identified on the basis of strong statistical evidence,
and further supported by linguistic and codicological findings).

In addition to bearing the pattern in their use of the signs, I found

in the use of rafe in Bible MSS is a complex issue which deserves sepa-
rate treatment.

A Further Analysis of the ‘Byzantine Triad’ of Features 165

that such MSS often exhibit trends in vowel sign interchange that
are reminiscent of, or may even reflect, forms of ‘Palestino-Tibe-
rian’ vocalisation. As this specific grouping of features appeared
to occur in MSS with palaeographies ranging between Italian,
Byzantine, and Levantine Oriental, this triad of features was des-
ignated ‘Byzantine’ (to describe the span of regions) (Arrant
2020, 515). The principal contribution of these studies was,
therefore, to conceive of manuscripts characterised by the Byz-
antine Triad of features as a distinctive ‘type’ of medieval Tibe-
rian Hebrew Bible.

However, a study devoted to the Byzantine Triad of fea-
tures alone, detailing its exact features within the corpus of He-
brew Bible manuscripts and exploring the impact this grouping
of features has upon the reading of the text as a coherent pattern,
has not been undertaken. Similarly, no attempt has yet been
made to engage with the codicological context(s) in which the
Triad appears and to consider its role in the reality of biblical
study and ritual use. Finally, Arrant (2021) identified three more
fragments that display the Byzantine Triad of features (two of
which appear to come from the same codex), which need to be
further contextualised with those published in Arrant (2020).

In the present article I seek to study the Byzantine Triad of
features on the basis of the broadest array of up-to-date evidence
available. I will describe the entire phenomenon in greater depth,
paying special attention to its linguistic function and impact upon
the text, contextualising the pattern of co-occurrence within its
codicological surroundings and suggesting ways in which it may

have functioned in practical use. I also discuss the terminology I

166 Arrant

have used to describe it, including a brief justification for the
term ‘Byzantine Triad’. All of the MSS studied separately in my
previous publication (Arrant 2020, especially 516-19; 2021) will
be considered together in context, with the rest of the data being
sourced from my PhD thesis (Arrant 2021). The MS fragments are
from Cambridge’s Taylor-Schechter and Lewis Gibson collections

and consist of Torah Bible codices on parchment.‘

1.0. The ‘Byzantine Triad’ of Features and Their

Purpose

It seems that, when the three aforementioned features co-occur
in a MS, they work in unison in an orthoepic manner to preserve
and reinforce the Tiberian Masoretic syllable structure of Biblical
Hebrew.” In this section I will describe the form and presentation
of each feature separately, and then analyse how they cooperate

to achieve such an effect.

1.1. Individual Feature Analysis

First, I will examine each element of the Byzantine Triad of fea-

tures alone and seek to understand its independent function.

* The eleven MSS currently identified as characterised by the full com-
plement of Byzantine Triad features are: T-S NS 21.6, T-S NS 248.5, T-
S NS 248.11, T-S NS 248.12, T-S NS 248.16, T-S NS 248.17, T-S Misc.
2.75, Or.1080 A.4.18, Or.1080 A.4.20, Or.1080 A.4.3, and T-S AS
64.238. The final three were identified as having the Byzantine Triad
in Arrant (2021). Or.1080 A.4.20 and T-S AS 64.238 seem to come from
the same codex.

° On the notion of orthoepy and its relevance to Hebrew Bible reading
traditions, see Khan (2018; 2020, I: 73-85, 99-105).

A Further Analysis of the ‘Byzantine Triad’ of Features 167

Since there is slight variation in the presentation of these signs
from codex to codex, I will also give details of such variations
and their significance for our understanding of the element’s
overall function. Note that throughout this article, all counts of
features are approximate: due to damage, an exact number for a
given feature cannot be relied upon. The counts do, however,
represent the majority of the MS texts and so are reliable as broad

indicators of the nature of the texts and their major trends.

1.1.1. ‘Dagesh’ in ’alef

In all manuscripts that show the Byzantine Triad of features, a
dot appears in ’alef, placed higher than the level of the vowels
(so as not to be mistaken for a hireq), between the midstroke and
left ‘foot’ of the ’alef. For example, 7x7) ‘and sees’ (Num. 21.8) in
T-S NS 21.6:

| a "e
ze
=
This sign occurs only in consonantal ’alef (i.e., ’alef with a vowel)

in the following manuscripts:
e T-S NS 21.6 (~39 identified occurrences), e.g., 9x7" ‘Is-
rael’ (Num. 20.28), jinx ‘Arnon’ (twice), 5x*9n3 ‘Nahali’el’

(twice);

168

Arrant

T-S NS 248.5 (~14 identified occurrences), e.g., nx direct
object marker (12 times); nnx ‘one’ (twice);

T-S NS 248.11 (~26 identified occurrences), e.g., nxynv”
‘from impurity’ (Lev. 16.19), nx direct object marker (15
times), Ix) ‘and after’ (Lev. 16.26);

T-S NS 248.12 (~95 identified occurrences), e.g., ni ‘and’
+ direct object marker (41 times); j44x ‘Aaron’ (three
times); Dx ‘and carried them’ (Lev. 10.5);

T-S NS 248.16 (~42 identified occurrences), e.g., 8") ‘rep-
resentatives of (Num. 26.9), nixa ‘hundreds’ (5 times),
Sixw? ‘to Saul’ (twice);

T-S NS 248.17 (~19 identified occurrences), e.g., Jina ‘the
female donkey’ (3 times), 1a7x ‘I speak’ (Num. 22.35), axia
‘Mo’ab’ (Num. 22.36);

T-S Misc.2.75 (~66 identified occurrences), e.g., wxd
‘to/for a man’ (7 times), o°W3x ‘men’ (twice), PMX» ‘to his
brother’ (three times);

Or.1080 A.4.18 (~48 occurrences), e.g., AXA ‘the Amo-
rite’ (6 times); nx direct object marker (15 times), 1xin
‘Mo’ab’ (5 times);

Or.1080 A.4.20 (~82 identified occurrences), e.g., TxA
‘very’ (twice), 0x ‘if (6 times), ixq’n ‘you (MPL) see’ (3
times), and its join T-S AS64.238 (~38 identified occur-
rences), e.g., oW3xn ‘the men’ (twice), DINW ‘and they (M)
showed them’ (Num. 13.26), irs ‘we saw’ (3 times);
Or.1080 A.4.3 (~28 identified occurrences), e.g., 1Wx ‘that’
(twice), ixvm ‘you (MPL) see’ (twice), pixn ‘the earth’ (3

times).

A Further Analysis of the ‘Byzantine Triad’ of Features 169

In general, this marking is consistent and regular; each time
a consonantal °alef appears in the text, it is marked with the sign.®
Therefore, it apparently does not serve to mark the occasional
-alef that readers might be prone to forget to pronounce.’ Fur-
thermore, as is evident from the examples above, the occurrence
of the sign is not conditioned by any specific positioning within
the word; it occurs in open syllables, closed syllables, and when
a vocalic ’alef is the first consonant in the word. Nor is the phe-
nomenon grammatically restricted; it does not occur only in
proper nouns, in prepositions, or with particular verbs, etc. It
seems, then, that this sign functions to mark, specifically, the con-
sonantal quality of vocalic ’alef and does so as a typical feature
of the diacritic system within these manuscripts. The apparent
function of this dagesh-like sign was to ensure that consonantal
-alef was not elided when the text was read aloud. The intention
was to preserve the sound (and, thus, the syllabification).

Further support for this position may be seen in the ten-
dency in these manuscripts to place a rafe on quiescent alef,
thereby explicitly marking that in such cases ’alef is not pro-

nounced as a consonant (Arrant 2020, 516-19).

° The exception in the present corpus is T-S NS 248.5, in which the sign
in question appears in only two words.

7 Such ‘utilitarian’ forms of non-standard vocalisation and diacritic use
do indeed appear in Geniza Bible manuscripts. They seem to function
almost like an aide memoire to help the reader pronounce only specific,
perhaps troublesome, words correctly; for a discussion of Bible MSS
with such utilitarian features, see chs 4 and 5 of Arrant (2021). This
phenomenon of a dagesh-like sign in consonantal ’alef is too regular for
such a function to be the case here.

170 Arrant

At this point one may question whether the sign should be
considered a dagesh or mappigq. On the basis of its consistent occur-
rence within the manuscripts on every consonantal ’alef, together
with the frequent simultaneous placement of rafe on quiescent
-alef, in a pattern that is not grammatically or semantically condi-
tioned, the sign is more akin to dagesh than mappiq. A comprehen-
sive discussion of this issue, which compiles relevant external evi-
dence, is found in Khan (2020, I:135-50), who convincingly shows
that grammarians of the time considered such a dot a dagesh forte,

doubling the ’alef to ensure its pronunciation when consonantal.®

1.1.2. Shewa on a Word-Final Guttural (‘ayin or het)

The second feature of the Byzantine Triad of features that appears
in all manuscripts,’ is the placing of a shewa on word-final, oth-
erwise unvocalised ‘ayin or het, e.g., n701 ‘and sent’ (Num. 21.32)
in Or.1080 A.4.18:

8 One should also note, however, that MSS like these (and various re-
lated MSS) also extend the use of the mappiq: “Mappiq is typically also
extended from word-final heh to word-initial and word-medial heh and
has the same function of marking the heh as consonantal” (Arrant 2020,
516).

° Since words ending with an unvocalised ‘ayin or het are comparatively
infrequent and because Geniza Bibles passages are often fragmentary,
it may be that there are more Bibles with the Triad than are analysed
in this article.

A Further Analysis of the ‘Byzantine Triad’ of Features 171

And on “ayin, e.g., pawrawsx ‘which swore’ (Num. 14.16) in
Or.1080 A.4.3:

Examples of this vocalisation in manuscripts exhibiting the Triad:!°

e T-S NS 21.6 (~3 identified occurrences), e.g., yaw ‘and he
listened’ (Num. 21.3), now» ‘and he sent’ (twice);

e T-S NS 248.11 (~7 identified occurrences), e.g., n7'71 ‘and
he will take’ (4 times), ndw> ‘to send’ (twice), yaw ‘seven’
(Lev. 16.19);

Note that T-S NS 248.5 is an outlier regarding this feature: it was
identified as a Byzantine Triad manuscript in Arrant (2020), but its
word-final shewa does not occur with a guttural. Instead, it occurs three
times in words ending in v-, for example, »y ‘upon it’ (Exod. 30.9). I
include it in this study because it was analysed in Arrant (2020). I view
the shewa here as having essentially the same function as shewa with a
word-final guttural: to signal to the reader that the final letter is conso-
nantal and that the syllable is closed. T-S NS 248.12 also has this feature
alongside shewa on word-final ‘ayin and het.

172 Arrant

e T-S NS 248.12 (~26 times, including on word final 1-, see
fn. 10), e.g., mp ‘and he took’ (Lev. 8.27), »337 ‘who
touches (M)’ (4 times);

e T-S NS 248.16 (~9 identified occurrences), e.g., Mp
‘Korah’ (twice), nynw) ‘to Shuthelah’ (twice), »23> ‘to
Bela’ (Num. 26.38);

e T-S NS 248.17 (~2 identified occurrences), e.g., nam ‘and
he offered’ (Num. 22.40), now ‘and he sent’ (Num. 22.40);

e T-S Misc.2.75 (~ 2 identified occurrences), e.g., mina ‘in
giving rest’ (Deut. 25.19, with patah under the yod and a
shewa under the het) and nawn ‘you (Ms) will forget’ (Deut.
25.19);

e Or.1080 A.4.18 (~5 identified occurrences), e.g., Np" ‘and
he took’ (twice), m7) ‘as an aroma’ (Num. 28.24), now ‘and
he sent’ (Num. 21.32; the dagesh in the lamed is non-stand-
ard as well);

e Or.1080 A.4.20 (once): xirnoe ‘forgive please’ (Num.
14.19);

e T-S AS 64.238 (~2 identified occurrences), e.g., N2w ‘he
sent’? (Num. 13.16), paw ‘seven’ (Num. 13.21);

e Or.1080 A.4.3 (~3 identified occurrences), e.g., vawi ‘he
swore’ (Num. 14.16), vwai ‘and transgression’ (Num.
14:18), xi-ndo ‘forgive please’ (Num. 14.19).

Occasionally, a naqdan confused furtive patah and a shewa meant
to close a syllable. For example, Or.1080 A.4.18 has an instance
where the naqdan placed a shewa where a furtive patah would be
expected: ni’1 (Num. 28.24). The manuscript does not show free

interchange of vocalic shewa and patah except in two places, both

A Further Analysis of the ‘Byzantine Triad’ of Features 173

instances where the naqdan substituted a patah or hatef patah for
shewa, e.g., 11p° for tty ‘Jazer’ (Num. 21.32); Aym@ for myn? ‘to
Jahaz’ (Num. 21.23). In the light of these cases, it appears that
the interchange between furtive patah and shewa here is, strictly
speaking, a case not of vowel interchange, but of shewa mistak-
enly placed under a guttural as if it closed a syllable. This inter-
pretation is strengthened by the placement of word-final shewa
below all other word-final gutturals that close the syllable within
this fragment.

This sign is used in an orthoepic manner, serving to pre-
serve the pronunciation of the gutturals and/or the proper syl-
labification of the text. When unvocalised ‘ayin or het appear at
the end of a word, they should invariably close the syllable. In
this phonetic environment, especially without a vowel such as
furtive patah, the guttural is vulnerable to elision from pronunci-
ation, due, it seems, to the weakening of gutturals in the reading
tradition of the scribe. This results in the loss of the sound of the
final consonant and creates an open syllable. Wherever there is a
word-final unvocalised guttural letter, by placing a shewa be-
neath that letter, the naqdan cues the reader to stop and close the

syllable, and, if possible, to try to pronounce the guttural.

1.1.3. Extended Use of dagesh forte

The third member of the Byzantine Triad of features is the place-
ment of dagesh forte in letters which are not otherwise geminated
according to the standard Tiberian Masoretic tradition. This sign

can occur in a range of letters, with the pattern appearing to be

174 Arrant

idiosyncratic, its extent determined by the judgement of the in-
dividual scribe. It occurs with high frequency in each manuscript
and has been studied in its own right in multiple pieces of schol-
arship (Morag 1959; Eldar 1978; Yeivin 1983; Khan 1991; 2017;
Blapp 2017).

In this section, I am interested in determining whether there
are meaningful details or patterns in the small variations of each
scribe’s use of extended dagesh. The aim is to identify conditioning
factors and to assess the degree of variation in usage between
MSS."! The factors in question are the letters that take extended
dagesh and their phonological context, i.e., where they appear in
the syllable, what sounds precede the geminated consonant, and
patterns of accentuation. In this section, I will first present and
describe the data, and thereafter engage with the scholarly discus-
sion surrounding the interpretation of this feature.

In the present corpus of eleven MSS, the majority of occur-
rences of extended dagesh occur at word-initial syllable onset.
Out of hundreds of cases of extended dagesh, only around 25 were
found to occur in the middle of a word. Of these word-medial

occurrences, the majority were located at syllable onset, after

™ In some cases, the examination involved a closer look than was pre-
viously possible; for damaged MSS I used microscopy at ~50x magnifi-
cation to help confirm or deny the possibility of the placement of dagesh.
This proved helpful, in that it allowed for the discovery of more features
than I originally found in my PhD research, and it also clarified points
where, to the naked eye, a dagesh may seem to be present, but in fact
the dot was a blemish on the writing surface and not ink. Due to this
and to manuscript damage, any counts of the occurrence of this ex-
tended dagesh are approximations and should not be taken as exact.

A Further Analysis of the ‘Byzantine Triad’ of Features 175

both silent and vocalic shewa, for example, nwpwpi ‘scales’ (T-S
NS 248.12, twice in the MS), unn>w ‘you (Ms) sent us’ (T-S AS
64.238, Num. 13.27). A small minority of these word-medial oc-
currences of extended dagesh were placed at the end of a syllable,
for example, in the samekh that closes the middle syllable in
maeini ‘and like its drink offering’ (Exod. 29.41, T-S NS 248.5)
and in ioi ‘and go out’ (Num. 14.25, Or.1080 A.4.20). These
dagesh signs within a word appear to be strategically placed where
consonants cluster at a syllable juncture so as to avoid the elision
of sounds at syllable onset or, rarely, syllable coda. They occur
mostly in the consonants tet, lamed, nun, and mem, and once in

Far more commonly, extended dagesh is placed in the first
consonant of a word, typically when that consonant is lamed,
mem, or nun, i.e., sonorant, especially nasal or labial, consonants.
Infrequently, the dagesh is placed in word-initial samekh, gof,
sade, and zayin. Extended dagesh at the beginning of a word ap-
pears to be more common when the final consonant of the pre-
ceding word is a sonorant (nasal or labial). It appears that the
dagesh serves to force the reader to stop and pronounce what is
effectively a doubled consonant, and to thereby distinguish be-
tween two similar (or identical) sounds.

Each manuscript, however, tends to have its own idiosyn-
cratic usage of this sign, which I will now explore. Generally, the
MSS discussed here tend to fall into two categories: those that use
extended dagesh at every opportunity (nearly every word-initial
lamed, mem, or nun), and those that are more selective (using ex-

tended dagesh at particular ‘problem points’ within the text). In

176 Arrant

Table 1 below, I summarise the main consonants in which word-
initial dagesh occurs, whether there is a trend for it to occur after
a disjunctive or conjunctive accent, and whether it occurs after a
word that ends in an open or closed syllable. Where a ‘slight pref-
erence’ is present, this indicates that the counts between options
are too close (nearly equal), and so a definite preference or cannot
be confidently stated given the condition of the MSS. See Table 1.

Table 1’s data reveal the following general trends. First, la-
med, mem, and nun are universally represented as taking extended
dagesh in every manuscript. Less regularly it appears in sibilants,
e.g., sade, samekh, and zayin. Second, with the exception of two
MSS (with nearly equal representation), extended dagesh is more
commonly written in word-initial consonants that follow a disjunc-
tive accent. This being the case, the number of times in which
word-initial extended dagesh is present following a conjunctive ac-
cent is sufficiently regular to argue that the type of accentuation
in the preceding word is not a major conditioning factor that trig-
gers the presence of extended dagesh in these MSS. The same mixed
picture holds for word-initial extended dagesh following open or
closed syllables. While there is a preference for closed syllables in
all but one manuscript, this preference is not strong enough for us
to definitively say that extended dagesh occurs characteristically
after a closed syllable and not after open syllables. Therefore, the
data appear to show that accentuation and syllable structure are

not determinative factors for the placement of the sign.'*

” Arrant (2020, 516ff.; 2021, 489) states that this dagesh occurs after a
vowelless consonant, summarising the current scholarly consensus on ex-
tended dagesh. The data here clarify the picture: the relevant dagesh tends

A Further Analysis of the ‘Byzantine Triad’ of Features 177

Table 1: Use of extended dagesh in word-initial consonants

Preference for occurring after

Classmark Bicone:

(description of porous pitee Se open/closed
ae J syllable

application) accent

Es ne 16 mainly: 5, n, 3 0c- disjunctive closed (definite:

(consistently, but not ex- f :

feusively) casionally: 0 (slight) ~21 to ~12)

Es Ne sininey mainly: 5, n; occa- disjunctive _ closed (definite:

(consistently, but not ex- . .

tensively) sionally: (slight) ~15 to ~4)

T-S NS 248.12

mainly: 5, n, ¥; oc- disjunctive

casionally: v, 1, p, (strong: ~51 loses Gsniie:

(extensively, to a lot of

letters, in a wide array ; to ~17) ~48 to ~22)
of contexts)

ie . Bias ._, Mainly: 5, n; once: disjanenve closed (definite:
(extensively to word-ini- as (strong: ~71 ~50 to ~32)
tial 5 and n) a to 13)

T-S NS 248.17

, : mainly: 5, n; once: disjunctive
(selectively; see discus- year J

closed (slight)

sion below) A (slight)
T-S NS 248.5 disjunctive: 1x
. . ; : ‘ closed: 1x
(selectively; see discus- twice: 5 conjunctive:
: open: 1x
sion below) 1x
T-S AS 64.238 + Or.1080 mainly: 4, n, 3; oc- ieinenas closed
A.4.20 (consistently, but — casionally: 1, 0, 0, J . (definite: ~19 to
. (slight)
not extensively) Xx, D ~3)
Or.1080 A.4.3 i ;
; ; disjunctive
(consistently, but not ex- mainly: 4, n, 3 oc- ce closed
tensively) casionally: 0, > (dehinitey se (slight)
y. y: DO, fF to ~13) &
Or.1080 A.4.18 ; nearly equal
: mainly: 5, n; occa- ae: :
(extensively) ‘ (disjunctive
sionally: 1 (1x), 3, , equal
bv, > (1x) ~25, conjunc-
ici tive ~22)
peal eiee may 2, 1 % OC- disjunctive closed (definite:
(consistently, but not ex- casionally: 1, v, 0, ;
‘ (slight) ~28 to ~11)
tensively) x

to occur after a vowelless consonant, but a significant number occur after
an open syllable, i.e., one that ends in a vowel.

178 Arrant

T-S Misc.2.75 , disjunctive =
: mainly: 5, 1, 3; oc- - open (definite:
(extensively) : (definite: ~25
casionally: 1, x to ~12) ~25 to ~14)

Given the data above, I would argue that extended dagesh
is primarily conditioned by consonant clusters involving lamed,
mem, and nun, when these letters are the second consonant in a
two-consonant cluster. Some manuscripts apply this feature ex-
tensively, so that nearly every word-initial lamed, mem, or nun
has a dagesh. Some apply it consistently, but not universally. But
most telling are those that apply the feature selectively. This is
enlightening, as we can see scribal choice at play in the use of
the sign. To demonstrate this, I will briefly discuss the two MSS
which apply extended dagesh only in certain phonological con-
texts: T-S NS 248.5 and T-S NS 248.17.

T-S NS 248.5

This manuscript has the smallest degree of usage of extended
dagesh. It occurs word-initially only twice and word-internally
once. The word-internal occurrence is 20121 for 130101 ‘and like
its drink offering’ (Exod. 29.41). The dagesh here appears to dis-
tinguish the samekh from the kaf and prevent the merging of the
sounds or the eliding of the samekh. The other two uses of ex-
tended dagesh in this manuscript are in the lameds in the phrase
> 172) wWipx ‘I will consecrate [them] to minister to me’ (Exod.
29.44). Again, the placement in consonant clusters reinforces the
distinction between sounds, but it is more noticeable in the sec-
ond occurrence, where the dagesh is placed in a lamed that occurs

after another sonorant (nun). It seems that the dagesh was placed

A Further Analysis of the ‘Byzantine Triad’ of Features 179

in locations that may have been tricky for a reader to pronounce

accurately when reading quickly.

T-S NS 248.17

While this manuscript includes this feature to a far greater degree
than T-S NS 248.5 (see above), its usage is still comparatively
infrequent relative to the other MSS studied in this article. I was
able to count only 19 instances of extended dagesh in this manu-
script, whereas in the other manuscripts the occurrences typically
trend up past 50 times. The instances where extended dagesh oc-
curs are either where there is vowel harmony (in the case of an
open syllable before the extended dagesh), or where there are
consonants between the two words which have points of articu-
lation that are close to each other (such as a dental following a

bilabial). The data can be broken down as follows:

e dagesh in word-initial alveolar lamed—occurs after mem (bi-
labial): p25 ov>a ‘Bil‘am to Balak’ (Num. 23.3) and ova
Tinxd ‘Bil‘am to the donkey’ (Num. 22.29); after taw and
resh:'° 79 nivy? ‘to do to you’ (Num. 22.30), X> AX", ‘and
he said “No.” (Num. 22.30); occurs after long vowels in an
open syllable (~6 times), e.g., 85 i> ‘and all of them [you
will] not [see]’ (Num. 23.13), 879 77x ‘[I sent] to you to

invite [you]’ (Num. 22.37); occurs after a guttural rein-

13 We do not know, of course, if this resh was realised as an alveolar or
uvular. In the example cited, it would have been pronounced as a uvular
rhotic in the standard Tiberian pronunciation tradition (Khan 2020,
1:223-34).

180 Arrant

forced with a shewa (once): ovya> nbv»,‘and he sent to
Bil‘am’ (Num. 22.40);

e dagesh in word-initial mem (bilabial)—occurs after resh and
lamed (three times): 77 1271 ‘and whatever [is revealed’
(Num. 23.3), aby ‘why’ (Num. 22.32), axia 7p ‘the city of
Moab’ (Num. 22.36); occurs after a labial (twice): Axin
~n79 ‘...Moab from the hills of...’ (Num. 23.7), 7 Ova

‘,..to Bil‘am “what...”’ (Num. 22.28); occurs after a dorsal

diphthong (once): 7Tiva %y ‘upon me, your whole life’
(Num. 22.30);

e dagesh in word-initial nun: occurs after a diphthong (once):
ax3 nm ‘[the angel of] the LorpD standing’ (Num. 22.31).

For the most part, these occurrences make sense if conceived in
terms of proximity in points of articulation: where there is a clus-
ter of coronal and labial consonants, a dagesh is placed to distin-
guish one consonant from the other.

The absence of this feature in other locations may serve to
explain such selectivity—that it is primarily difficult consonant
clusters which trigger the placement of extended dagesh in select
manuscripts. In MSS that use this feature comprehensively, it ap-
pears that the usage has become systematic in its application
throughout the whole of the text, particularly for lamed, mem,

and nun. I would argue that for such comprehensive occurrences,

A Further Analysis of the ‘Byzantine Triad’ of Features 181

the original intention is still the same, but the feature spread to
all occurrences as a normalising function of the diacritic.'

These data offer some modifications to the scholarly discus-
sion on extended dagesh. Yeivin (1983, 297) and Khan (2017, 267)
discuss extended dagesh at word-initial syllable onset as typically
occurring after a closed syllable, and at word-medial syllable onset
as typically following silent shewa (Khan 2017, 267) and/or dif-
ferentiating between two similar letters (Yeivin 1983, 297). Khan
(2017, 267-69) describes cases of extended dagesh also occurring
in a dehiq structure, i.e., a dagesh placed word-initially following a
word with a conjunctive accent ending in an unstressed open syl-
lable. Moreover, scholars debate the phonetic function of the sign,
with Morag arguing that these signs break syllable boundaries, and
Eldar (1978, 125-43) terming the sign 15 w37 ‘separative
dagesh’, and Khan concluding that the sign is a dagesh forte func-
tioning orthoepically to distinguish syllable and consonant divi-
sions.

These astute observations are not contradicted by the Byz-
antine Triad MSS presented here. I would argue, however, that
the data show some additional trends that may (depending on
further research) prove unique to MSS characterised by the Byz-
antine Triad of features. First, in these latter MSS, extended
dagesh does not show a strong preference for occurring after a
closed syllable, but rather, often occurs after open syllables.

Moreover, this goes beyond a classic dehiq structure, in that after

Khan (2017, 267-68) gives an excellent analysis of the phonological
impact of this sign to distinguish the syllables and to reinforce the pro-
nunciation of the second element in a syllable division.

182 Arrant

an open syllable, the sign can occur regardless of whether the
preceding accent is conjunctive or disjunctive and regardless of
whether the preceding syllable is stressed or unstressed. There-
fore, while MSS seem to show a slight preference for the sign
after disjunctive accents and closed syllables, this is by no means
the typical presentation of the feature. Indeed, it appears that the
majority of these MSS take extended dagesh according to the con-
ditioning factors discussed in scholarship, but further extend it,
placing it in any word-initial lamed, mem, or nun by default. In
the case of those MSS where the feature is selectively placed, the
primary conditioning factor is the desire to ensure careful read-
ing at difficult consonant clusters across words, whatever the pre-
ceding syllable’s status or accentuation. My claim, therefore, is
that these MSS represent one type of extended dagesh within the

overall phenomenon.'®

A further observation that has come to light in one of the
MSS supports this claim. T-S NS 21.6 adds a paseq between the last
two words of 773 5X77 771" ‘and Israel made a vow’ (Num. 21.2):

This sign is not attested in BHS/L at this location, but it is clearly
a paseq, since the sign is identical to other instances of paseq

within the MS. Here its function appears to be to enhance further

8 It is to be noted that Yeivin (1983) and Khan (2017) both
acknowledge that within the patterns they describe, there are many
forms of variation and exceptions. The case of these MSS appears to be
such an instance.

A Further Analysis of the ‘Byzantine Triad’ of Features 183

the distinction between the lamed and the nun, forcing the reader
to stop and pronounce the words separately. This instance of non-
standard accentuation clearly correlates with other orthoepic

functions of the dagesh in the oral reading of the text.

1.2. The Features in Tandem

We have seen in the above sections the individual presentation
of each component of the Byzantine Triad of features. As these
features co-occur in the manuscripts, however, they should be
seen as complementing one another. In this sense, the Triad ap-
pears to be an attempt to preserve accuracy in the reading, par-
ticularly by reinforcing correct syllabification. The insertion of a
dagesh preserves the pronunciation of vocalic ’alef by making the
reader pause, and thereby ensures that the syllable remains in-
tact. By marking word-final unvocalised gutturals, the shewa re-
minds the reader that the syllable is closed (and also helps to
preserve the pronunciation, however weak, of the guttural). Fi-
nally, by placing a dagesh in consonant clusters that are particu-
larly vulnerable, the syllabification both within words and be-
tween words is preserved by signalling to the reader to pro-
nounce with added force consonants at risk of being slurred over
during reading.

It is noteworthy that the vocalisers of these texts made cre-
ative use of the Tiberian system of signs to encourage correct syl-
labification through extending the rules of their placement. None
of the signs are technically used incorrectly with regards to its
essential function: the dagesh forte sign is used here, as it is in

masoretic codices, to geminate the consonant. Likewise, silent

184 Arrant

shewa here closes syllables according to its standard function. In-
deed, even dagesh in ’alef is attested in masoretic codices, some-
thing Khan discusses extensively (2020, I:135-50). What appears
to have happened in Bibles with the Byzantine Triad of features
is that the function of the signs is used creatively, with non-stand-

ard placement, to promote a more careful, masoretic reading.'®

1.3. Vowel Sign Interchange Patterns Associated with

the Byzantine Triad of Features

Manuscripts with the Byzantine Triad of features described above
were found in Arrant (2020; 2021) to have important distinctions
when compared with a large number of Bibles from the same
corpus. A characteristic phenomenon in many ‘near-model’ and
‘common’ Bibles from the Cairo Geniza is the presence of a diz-
zying array of vowel sign interchanges in seemingly idiosyncratic
ways from manuscript to manuscript. Especially in Arrant
(2021), it was established that such interchanges are neither ran-

dom nor meaningless. Different patterns of vowel sign inter-

© There are a few other non-standard features that characterise these
manuscripts, but not at the consistent level of a pattern: there are occa-
sional irregularities in begedkefet notation; many of these Byzantine
Triad MSS (along with other MSS with close palaeographies) place the
shin dot within the shin (or even double the dot, with a dot inside the
shin and a dot atop the shin); at times the dagesh in the zayin of n1n ‘this’
is dropped. These features deserve further exploration outside of the
current study.

A Further Analysis of the ‘Byzantine Triad’ of Features 185

change correlate statistically with codicological features, re-
gional distinctions in palaeography, and each other in distinct
patterns that are linguistically meaningful.’

Arrant (2020) gives an overview of the Triad features and
vowel sign interchanges of eight MSS discussed in the present
study: T-S NS 21.6, T-S NS 248.5, T-S NS 248.11, T-S NS 248.12,
T-S NS 248.16, T-S NS 248.17, T-S Misc.2.75, and Or.1080
A.4.18. As the present study examines an additional three MSS—
T-S AS 64.238, Or.1080 A.4.20, and Or.1080 A.4.3—it provides
an updated picture of the vowel sign interchange data available
for Bibles with the Byzantine Triad of features.

Arrant (2020, 514) notes that the MSS analysed in that
study presented interchange patterns fitting Schema Patterns X,

Y, 1, and 1a. To review:

e Pattern X: MSS with this pattern have regular interchange
of shewa (usually vocalic) with patah, indicating that the
MSS belong to a tradition which pronounced shewa as [a].

e Pattern Y: MSS with this pattern feature a three-way inter-
change of shewa, hireq, and sere, probably reflecting raising
of the quality of vocalic shewa.

e Pattern 1: in MSS with this pattern, patah and qames freely
interchange, on the one hand, and sere and segol freely in-

terchange, on the other. There is no exchange between a-

” The statistical backing for this correlation is strong; approximately
409 codices (out of around 1400 codices comprising 1851 leaves) in the
corpus of Arrant (2020; 2021) had such ‘non-standard’ Tiberian vocali-
sation and, so, can be considered sufficient for a representative sample.
Cf. Arrant (2021, 29-63) for the statistical methodology.

186 Arrant

and e-class vowels. This effectively reduces the vowel in-
ventory to five, with single /a/ and /e/ vowels.

e Pattern 1a: related to, but unlike Pattern 1, in this pattern
qames and patah remain distinct (so that there are two re-
alisations a-class vowels), but sere and segol have merged
into a single e-class vowel (as indicated by their free inter-

change throughout the manuscript in question).

In Arrant (2020; 2021), patterns X and Y were described as ‘no-
tational’ interchanges, where vocalic shewa was simply replaced
with the vowel sign of the equivalent vocalic quality, i.e., patah,
sere, or hireq, depending on the pattern.'* Patterns 1 and 1a con-
sist of true phonological interchanges of vowels, reflecting a
vowel system in the Hebrew pronunciation of the MSS that dif-
fered from that of the standard Tiberian pronunciation. Such an
inventory has been identified as ‘Palestinian’ in quality
(Heijmans 2013). Thus, it appears that the vocalisers of these
MSS sought to preserve syllable structure according to the rules
of the Tiberian Masoretic system, though their individual phono-
logical profile differed along the trend of realising vocalic shewa
as a raised vowel, and in some MSS, of reducing the vocalic in-
ventory to that of a five-vowel system of pronunciation, which is

characteristic of the Palestinian pronunciation tradition.

18 Tt is important to distinguish the two: the interchange between sere and
segol is a true vowel interchange, where the two /e/ vowels have merged
into one pronunciation. The interchange of high vowels with shewa—
which in these Bibles had an /e/ realisation (rather than the /a/ realisa-
tion of Pattern X)—is a notational, rather than a phonetic, distinction.

A Further Analysis of the ‘Byzantine Triad’ of Features 187

The MSS analysed in Arrant (2020) were of professional
codicological quality. Their diacritic differences show striving to-
wards the preservation of the syllabification and pronunciation
of the consonants, while their vowel sign interchanges may re-
flect the realities of Hebrew pronunciation in the region(s) in
which they were copied and used. As mentioned above, they
range in palaeography from Italian to Levantine Oriental, and
their vowel sign interchange (with its Palestinian Hebrew associ-
ations) appears to go in hand with such regional designations. In
this study we have added three additional fragments (two of
which are related) which appear to be codicologically and palae-
ographically similar to the MSS studied in Arrant (2020), with
the exception of one (discussed below). However, they are less
formal than the MSS studied in Arrant (2020); for example, they
lack masora, and one is smaller and has only one column. The
three fragments here (T-S AS 64.238, Or.1080 A.4.20, and
Or.1080 A.4.3), therefore, represent a slight expansion in terms
of codicological features from the originally identified Byzantine
Triad group. Here, therefore, we explore whether there is a
slightly wider profile of vocalic interchange present alongside a

slightly wider codicological and palaeographic range.!°

'° One may notice here my reticence to mention palaeographic dating.
I am hesitant to ascribe dates to manuscript fragments, mainly because
both script styles and linguistic features can become fossilised and per-
sist for quite some time. While I do give tentative dating estimates be-
low, it is more meaningful here, in my opinion, to show relationships
between objectively verifiable features (such as similarity of script or
vowel sign interchange), than to make an argument for trends based on

188 Arrant

Table 2 summarises the vowel interchanges in the 2020
case study contrasted with those of the three additional manu-

scripts included in this study.

Table 2: Comparative summary of vowel interchanges in MSS studied
in Arrant (2020; 2021); NI= notational interchange; VSI= vowel sign
interchange reflecting deviation from Tiberian pronunciation; numbers
in parentheses indicate count of occurrences

2020 Case Study ‘Near-Model’ Torahs

Or.1080 A.4.18: Patterns X, Y, 1

NI: patah/hatef patah and (silent) shewa (1)

VSI: patah for games (2); games for patah (9); segol for sere (2); sere for segol (1);
sere for hireq (1); sere for patah* (1)

*Note minimal interchange between sere and patah, violating Pattern 1.

T-S NS 248.11: Patterns X, 1

NI: patah for hatef patah (5)

VSI: segol for hatef patah* (1); shewa for patah (1); segol for sere (1); sere for segol
(1); patah for games (1)

*wxa for wx (Lev. 16.15) appears to be a unique case of vowel harmony;
every other instance of 1Wx in the MS is vocalised with patah instead of hatef
patah (and no patah-segol interchange).

T-S NS 248.17: Pattern 1?*

VSI: hireq for patah (1): 789m for 7890 (Num. 22.35); games for patah (1) nam
for nari (Num. 22.40)

*This manuscript is an outlier; unlike most of the other manuscripts, it does not
have any notational interchange, and its vowel interchanges are very minimal.
T-S NS 21.6: Patterns 1, Y*

VSI: patah for games (1); shewa for hireq (5); sere for segol (2)

*Like T-S NS 248.17, this manuscript has no notational interchange.

T-S Misc.2.75: Patterns X, 1a”°

VSI: shewa for games (1); sere for segol (1)

palaeographic dating (though this does not reduce my estimation of the
usefulness of palaeographic dating in other scientific contexts).

° Contrary to my 2020 article that identified it erroneously as Pattern
Y, the MS does not interchange shewa, hireq, and segol.

A Further Analysis of the ‘Byzantine Triad’ of Features 189

T-S NS 248.5: Patterns X, 2a

NI: hatef patah for patah (1); hatef games for qames (1)

VSI: games for patah (1); games for segol (1); segol for sere (1); shewa for patah
(4); shewa for segol (1)

T-S NS 248.12: Pattern Y?

VSI: shewa for segol (2)

T-S NS 248.16: Patterns Y

NI: hatef patah for shewa (1); hatef qames for qames (1)

VSI: hireq for shewa (1); hireq for shureq (1), patah for games* (1)

*Some minimal indication of Pattern 1, but incomplete.

2021 PhD Torahs (with basic codicological information)

T-S AS64.238 (+ join with OR.1080 A.4.20; two-column parchment codex,
portrait format,”! no Masoretic notes): Pattern X

NI: patah for hatef patah (6); patah for shewa (1); shewa for furtive patah (1);
shewa for hatef patah (4)

VSI: games for patah* (1)

**Some minimal indication of Pattern 1, but incomplete.

Or.1080 A.4.20 (two-column parchment codex, portrait format, no Masoretic
notes): Patterns X, 2a

NI: hatef patah for patah (1); patah for hatef patah (9); shewa for patah (2); segol
for hatef segol (1); shureq for qubbus (2); shureq for shewa* (1)

VSI: patah for segol (1); qames for patah (1)

*The shureq for shewa occurs once, on ivy for ipa (Num. 14.24).

Or.1080 A.4.3 (1 column parchment codex, landscape format, no Masoretic notes)
Patterns X, 2b

NI: hatef patah for patah (6); shewa for furtive patah (2); shewa for hatef patah (3)
VSI: patah for segol (1); segol for sere (1); sere for segol (2)

The table above indicates a trend of interchange consistent with
Patterns X, Y, 1, and 1a. A minority of MSS in the 2020 case study
have vowel sign interchange patterns that are typically seen in
Bibles where the vowel inventory is reduced under (presumably)
Arabic phonological influence (specifically, patterns 2a and 2b)
(see Arrant 2021, 157ff.). Two of the additional three MSS in-
cluded in the present study also have 2a/2b.

1 Portrait format = length (of a page) > width; landscape format:
width > length.

190 Arrant

The defining feature of these MSS as a whole seems to be
the rarity of interchange phenomena; when vowel sign inter-
changes occur, they are not pervasive, but usually occur only
once or twice. Notational interchanges tend to happen with
greater frequency than vowel sign interchanges. There seems to
be no meaningful difference between the MSS in the 2020 case
study and the additional three MSS in terms of vowel sign or no-
tational interchanges.

Therefore, the profile of Byzantine Triad Bibles seems to be

a tendency for:

e frequent notational interchange between hatef vowels,
patah, and shewa;

e relative infrequency of vowel sign interchanges indicative
of ‘Palestinian’ Hebrew phonology (Patterns Y, 1, 1a);

e outliers with very minimal interchanges indicative of a re-

duced vowel inventory (in Patterns 2a and 2b).

This picture is consistent, then, with a general ‘Palestino-Tiberi-

an’ association of linguistic phenomena regarding vocalisation.

2.0. The ‘Book-Type’ of the Byzantine Triad of

Features

As Arrant (2020; 2021) has demonstrated that vocalisation and
codicological features are mutually informative and that patterns
between vocalisation and codicology often correlate, only a few
cursory observations about the codicology of Byzantine Triad Bi-
bles are necessary. Arrant (2020) contextualised the ‘near-model’
Byzantine Triad MSS (two-three columns, on parchment, with

partial masoretic notes) among other ‘near-model’ Bibles lacking

A Further Analysis of the ‘Byzantine Triad’ of Features 191

the Byzantine Triad. Arrant (2021) dealt with the three addi-
tional, ‘non-model’ Byzantine Triad MSS within the context of
other Bibles without the Triad, but with a similar codicology.”
Because Bible codices with the Byzantine Triad of features share
codicological styles with Bibles that lack these features, they are
not completely codicologically distinct from Bibles with standard
and non-standard Tiberian vocalisation. However, their codicil-
ogy is still informative as to the contexts in which they were used.
In this section I will discuss the codicological relationship be-
tween ‘near-model’ and ‘non-model’ Bibles with the Triad fea-
tures, give observations on their palaeographic range, and make

inferences about their practical function(s).

2.1. Near-Model Byzantine Triad Codices

The codices examined in Arrant (2020) were ‘near-model’: all are
written on parchment and have partial masoretic notes. The first
observation of note is that all but one of these Byzantine Triad
codices (T-S NS 248.12) has two columns rather than three and,
so, are by default smaller and less grandiose than full, exemplary,
three-column Masoretic Bibles. They are ruled, tend to be pricked
on the outside margin (T-S Misc.2.75 is pricked on both margins,
while Or.1080 A.4.20 and T-S AS 64.238 are not pricked at all).
They are plain, with no illumination or ornate decoration. Their
script is smaller than that seen in the grandiose Oriental exem-
plary codices (though this is expected, as smaller script is typical

of Italian, Byzantine, and Southwestern Oriental script types). All

2 See Arrant (2021, chs. 3-4) for a contextualised discussion of Byzan-
tine ‘Trio’ Bibles within the larger corpus.

192 Arrant

of them have masora parva, but not masora magna. They range in
size from 18.4-31.7 cm long x 15.65-25.3 cm wide, i.e., on the
smaller side of Bible codices.*? They all have a portrait format
(length greater than width). They have a range of 19-27 lines per
page. Indeed, the combination of their general minimalist ap-
pearance, skilled writing, careful vocalisation, and small size
seems to indicate that these codices were carefully written, yet
intended for practical use.

They have the following palaeographic and codicological

ranges:

e T-S Misc.2.75: two columns, 26.6 x 22.7 cm, 26 lines, Ital-
ian, 12th or 13th c.

e T-S NS 21.6: two columns, 21 x 19.8 cm, 20 lines, Italian-
Byzantine, probably 12th c.

e Or.1080 A.4.18: two columns, 18.4 x 15.6 cm, 19 lines,
Italian, 12th c.

e T-S NS 248.5: two columns, 22.2 x 9.2 cm, 21 lines, Levan-
tine Oriental-Byzantine (from the Levant; Syria-Palestine,
but not an earlier calligraphic hand such as seen in the
Aleppo Codex; appears to have some scattered ‘Byzantine’
features).

e T-S NS 248.11: two columns, 21.4 x 19 cm, 23 lines, Levan-

tine Oriental-Byzantine.

3 For example, Arrant (2020) discusses a type of Bible there termed
‘Large Monumental Levantine Codex’, which ranges in size from 35-
38.2 cm long and 32-35 cm wide, and had 25-30 lines. Multiple groups
are discussed in Arrant (2020; 2021), all of which are significantly
larger than the eleven MSS studied here.

A Further Analysis of the ‘Byzantine Triad’ of Features 193

Figure 1: Near-Model Byzantine Triad Codices: T-S Misc.2.75r (top left),
T-S NS 248.11r (top right), Or.1080 A.4.18v (bottom)

Or IOBO.A. de AB.
oo
————S——

194 Arrant

e T-S NS 248.17: two columns, 24.3 x 19.3 cm, 23 lines, Le-
vantine Oriental-Byzantine.

e T-S NS 248.12: two columns, 31.7 x 25.3 cm, 27 lines, Le-
vantine Oriental-Byzantine.

e T-S NS 248.16: two columns, 21.7 x 23.1 cm, 20 lines, Ital-

ian-Byzantine.

2.2. ‘Non-Model’ Byzantine Triad Codices

The three additional codices studied here were analysed in Arrant
(2021). As the thesis did not study any ‘near-model’ codices,
these are slightly distinct from the group above. Two of the frag-
ments are similar to the above ‘near-model’ group in terms of

size, column number, and number of lines:

e Or.1080 A.4.20: two columns. 22.2 x 19.3 cm, 23 lines, Le-
vantine Oriental to Byzantine
e T-S AS 64.238: two columns. 20.7 x 18.9 cm, 23 lines, Le-

vantine Oriental to Byzantine

Further analysis of the handwriting and the fact that they have
consecutive passages (T-S AS 64.238 has Num. 13.7-14.6,
Or.1080 A.4.20 Num. 14.7-35) indicates that they are in fact two
leaves from the same Bible codex. Visually, they are nearly iden-
tical to the MSS of the above group, except that they lack maso-
retic notation.

Our final Bible, Or.1080 A.4.3, is unique. It is a single-col-
umn parchment codex in landscape format (12.5 cm long x 16.1
cm wide). It appears to have Italian (circa 12th c.) palaeography.

Unlike the other two Bibles here or those in the near-model

A Further Analysis of the ‘Byzantine Triad’ of Features 195

Figure 2: ‘Non-Model’ Byzantine Triad Codices: Or.1080 A.4.20r (top
left), T-S AS 64.238r (top right), and Or.1080 A.4.3v (bottom)

Mths wie WER RAT ON errs samy iabs yor 1 Ope
“Ba ea Mamebegn 0 a sa Lsasmece eR

Nis apheen paver er (a othe be ees ¥

pe Srna Y

ey ns mas oo SS* Bagram y
sen SMA Ad MD NEY ogee gas
seen ae svgenr: Br LOW ITTIN 36:
ben pave Naa 9 as Racobae

v7 ac

Shee ed as V3 S53

196 Arrant

group, it has only 15 lines per page. Its compact size, small writ-
ing, wide margins, and format make it appear informal, but the
quality of the script is exquisite.

It appears that, with the exception of Or.1080 A.4.3 (be-
cause of its landscape format), the Bibles with the Byzantine
Triad of features are very similar in appearance and come from a
restricted range of palaeographic regions.

Given their features, what can we infer about their pur-
pose? We must remember the careful vocalisation which seeks to
ensure correct syllabification and pronunciation of gutturals.
Such usage of the signs appears to indicate that these manuscripts
were read aloud, as the signs themselves have no independent
grammatical meaning except to ensure the prescribed syllable
pronunciation. One could read the text silently without these
signs and still grasp the correct grammar and understand the con-
tent. When read aloud these signs fulfil their purpose.

The small size of these Bibles indicates that they were not
grandiose endeavours and were not meant to be perfect speci-
mens of an elaborate, beautiful, masoretic work. They are care-
fully made and vocalised, yet still have an informal character.
While we can only guess as to whether they were read at home,
or in the synagogue, or both, we can clearly see that they were
to be studied and read aloud. The rewriting on T-S AS 64.238,
indicates that it was used for quite some time and may have
needed repair. I propose that they may have been made for the
purpose of study and preparation for reading the Torah aloud in
a didactic setting (whether a synagogue service or at home) and,

therefore, are ‘personal’, yet ritualistic. They are small enough to

A Further Analysis of the ‘Byzantine Triad’ of Features 197

be easily carried and held, yet written with sufficient care to be

professional, skilful, aesthetically pleasing, and textually reliable.

3.0. Conclusions

This article has assessed eleven Bible codices from the Geniza
which are strongly similar on linguistic and codicological
grounds. In conclusion, I will briefly discuss their diacritic, codi-
cological, and palaeographic connections to other kinds of Bibles,
and finally, the terminology we may choose to use to describe
them.

3.1. Vocalisation: Connections to ‘Extended Tiberian’

These Bibles have many of the features that scholars have come
to associate with ‘Extended Tiberian’ vocalisation, to the degree
that they may be considered an integral part of that phenomenon.
However, they are a distinctive subgroup of the extended Tiberi-
an tradition in their manifestation of particular features. These
include their regular application of shewa to word-final gutturals,
the further extension of what we typically consider ‘extended
dagesh’ to cover nearly all instances where a sonorant, especially
nasal or labial, consonant begins a word (regardless of the sylla-
ble or accentuation status of the preceding word), and their
placement of dagesh in consonantal ’alef. This subtype of Ex-
tended Tiberian is closely related to Palestino-Tiberian, with its
close association with Palestinian pronunciation. Further re-
search may be necessary to distinguish any other patterns within
such MSS.

198 Arrant

3.2. Codicology and Palaeography: Connections from

the Levant to Italy

The Bibles in this study have been assessed codicologically in
terms of functional implications. We have also noted that they
seem to represent a palaeographic range from the Levant up to
Italy. The book type represented has connections to other Bibles
studied in Arrant (2020, esp. 536; 2021, esp. 220) from Italy, and
the handwriting and extended vocalisation is very similar to that
of Codex Reuchlinianus, for example. Though they are few, I
would argue that such coherence is potentially evidence of
scribal connections through regions, from the Levant, through

Syria and Greece, into Italy, and then up into Ashkenaz.”*

3.3. A ‘Byzantine’ Triad of Features?

The final aspect of these Bibles that I will address here is the ad-
jective which I have used to describe them: ‘Byzantine’. It is sim-
ultaneously accurate, and in some ways also misleading. It is true
that the script type of these Bibles ranges from Italian, to Italian-
Byzantine, to Levantine Oriental-Byzantine. The original motiva-
tion for calling them Byzantine was because this range of repre-
sentation covers the region of Western Asia Minor and slightly

beyond, up to Italy and down to Palestine. Therefore, ‘Italian-

4 Note that Khan and other scholars also trace the features of Extended
Tiberian from Italy up into Ashkenaz (cf. Khan, 2017, 270). The im-
portance of this observation for our understanding of the history of the
transmission of the Hebrew Bible in the medieval period will be ex-
plored further in the future expanded and updated publication of my
PhD thesis.

A Further Analysis of the ‘Byzantine Triad’ of Features 199

Levantine’ is also an accurate descriptor. However, this is only a
palaeographic factor. Their unique pattern of vocalisation also
can be described in ways other than three chosen features; for
example, these three features do not cover the extensive place-
ment of rafe, or the vowel sign interchanges involved.” Classifi-
cation on the bases of these three features also does not indicate
the inherent connections these Bibles have to the Extended Tiber-
ian tradition. Yet the Byzantine Triad of features was specifically
chosen, because Bibles with all three features appear very simi-
lar; there are many other kinds of Bibles which have one or two
features of the Triad, but differ from these both codicologically
and textually. In short, there is no single term perfectly apt for
describing these Bibles in all their nuances. I submit that for the
time being, Byzantine Triad, or even Italian-Levantine Triad,
must suffice. However, it is possible that with further research,
additional manuscript fragments with these features will surface,
and further analysis on other aspects may turn up more suitable
descriptors. This conclusion, therefore, is certainly not the final

word on these fascinating Bible manuscripts.

References

Arrant, Estara J. 2020. ‘An Exploratory Typology of Near-Model
and Non-Standard Tiberian Torah Manuscripts from the

Cairo Genizah’. In Studies in Semitic Vocalisation, edited by

° Nor does it deal with the diverse features mentioned in n. 16, though
these do not appear to occur with measurable regularity.

200 Arrant

Aaron D. Hornkohl and Geoffrey Khan, 467-548. Cam-
bridge: University of Cambridge and Open Book Publishers.

https://doi.org/10.11647/OBP.0207.07
. 2021. ‘A Codicological and Linguistic Typology of Com-

mon Torah Codices from the Cairo Genizah’. PhD disserta-
tion, University of Cambridge.

Blapp, Samuel. 2017. ‘The Non-Standard Tiberian Hebrew Lan-
guage Tradition According to Bible Manuscripts from the
Cairo Genizah’. PhD dissertation, University of Cambridge.

Diez Macho, Alejandro. 1956. ‘Un Manuscrito Hebreo Protoma-
soretico y Nueva Teoria acerca de los Llamados MSS. Ben-
Naftali’. Estudios biblicos 15: 187-213.

.1963. ‘A New List of so-called “Ben Naftali” Manuscripts:

Preceded by an Inquiry into the True Nature of These Man-
uscripts.’ In Hebrew and Semitic Studies presented to Godfrey
Rolles Driver in Celebration of his Seventieth Birthday, 16-52.
Oxford: Clarendon Press.

Eldar, Ilan. 1978. The Hebrew Language Tradition in Medieval Ash-
kenaz (ca. 940-1350 CE). Jerusalem: Magnes Press, The He-
brew University of Jerusalem.

Heijmans, Shai. 2013. ‘Vocalization, Palestinian’. In Encyclopedia
of Hebrew Language and Linguistics, edited by Geoffrey Khan
et al., III:964-67. Leiden: Brill.

Khan, Geoffrey. 1991. ‘The Syllabic Nature of Tiberian Hebrew
Vocalization’. In Semitic Studies: In Honor of Wolf Leslau on
the Occasion of His Eighty-fifth Birthday, November 14th,
1991, edited by Alan S. Kaye, 850-65. Wiesbaden: Har-

rassowitz.

A Further Analysis of the ‘Byzantine Triad’ of Features 201

.2017. ‘The Background of the So-called “Extended Tibe-
rian” Vocalization of Hebrew’. Journal of Near Eastern Stud-
ies 76/2: 265-73.

.2018. ‘Orthoepy in the Tiberian Reading Tradition of the

Hebrew Bible and Its Historical Roots in the Second Temple
Period’. Vetus Testamentum 68: 1-24.
. 2020. The Tiberian Pronunciation Tradition of Biblical He-

brew. 2 vols. Cambridge Semitic Languages and Cultures 1.
Cambridge: University of Cambridge and Open Book Pub-
lishers. https://doi.org/10.11647/OBP.0194.01

Morag, Shelomo. 1959 ‘The Vocalization of Codex Reuchlinianus:
Is the “Pre-Masoretic” Bible Pre-Masoretic?’. Journal of Se-
mitic Studies 4/3: 216-37.

Yeivin, Israel. 1983. ‘Mashma‘ut Siman ha-Dagesh ba-Niqud ha-
Tavrani ha-Murhav.’ In Hebrew Language Studies Presented to
Professor Zeev Ben-Hayyim, edited by Moshe Bar-Asher et
al., 293-307. Jerusalem: Magnes Press, The Hebrew Uni-

versity of Jerusalem.

HEBREW VOCALISATION SIGNS IN
KARAITE TRANSCRIPTIONS OF
THE HEBREW BIBLE INTO
ARABIC SCRIPT

Geoffrey Khan

1.0. The Karaite Transcriptions

In the 10th and 11th centuries CE many Karaite scribes in the
Middle East used Arabic script to write not only the Arabic lan-
guage, but also the Hebrew language. Such Hebrew texts in Ara-
bic transcription were predominantly Hebrew Bible texts. These
were sometimes written as separate manuscripts containing con-
tinuous Bible texts. Some manuscripts in Arabic script contain
collections of biblical verses for liturgical purposes. Arabic tran-
scriptions of verses from the Hebrew Biblical or individual Bibli-
cal Hebrew words were in many cases embedded within Karaite
Arabic works, mainly of an exegetical nature, but also in works
of other intellectual genres. Several Karaite Arabic works also
contain Arabic transcriptions of extracts from Rabbinic Hebrew
texts (Tirosh-Becker 2011). The Karaites transcribed into Arabic
script only texts with an oral reading tradition, as was the case
with the Hebrew Bible and rabbinic texts in the Middle Ages. The
transcriptions reflect, in principle, these oral traditions. It is for
this reason that their transcription of the Hebrew Bible represents

204 Khan

the gere (the orally transmitted reading tradition of the text) ra-
ther than the ketiv (the written tradition). Other types of Hebrew
text that were written by Karaites during the Middle Ages with-
out an oral tradition, e.g., documents, commentaries, law books,
were always written in Hebrew script (Khan 1992).

Most of the known manuscripts containing Karaite tran-
scriptions of Hebrew into Arabic script are found in the British
Library (Khan 1993), the Firkovitch collections of the National
Library of Russia in St. Petersburg (Harviainen 1993), and in the
Cairo Geniza collections (Khan 1990). These manuscripts ema-
nate from Palestinian circles of Karaites or Karaites in Egypt who
had migrated to Egypt from Palestine after the capture of Jerusa-
lem by the Crusaders in 1099. The majority of them were written
in the 10th and 11th centuries.

Most of the transcriptions of Biblical Hebrew reflect the Ti-
berian reading tradition or an attempt to reflect this tradition.

The Tiberian pronunciation tradition of Biblical Hebrew
was regarded as prestigious and authoritative in the medieval
Middle East. It is likely that the authoritativeness of the Tiberian
tradition had its roots primarily in its association with the Pales-
tinian Yeshiva ‘Academy’, the central body of Jewish communal
authority in Palestine, which was based in Tiberias from late an-
tiquity until the Middle Ages. The Masoretes were closely associ-
ated with the Palestinian Yeshiva (Khan 2020b, I:86). Due to its
authority and prestige, the Tiberian pronunciation was the ideal
target in the oral reading of the Bible in communities. In such
situations, outside the inner circles of the masoretic masters of
Tiberias, there was always a risk that the ideal target would have
been missed, resulting in an imperfect performance of the Tibe-

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 205

rian tradition. In a previous paper (Khan 2020a), I discussed var-
ious aspects of the imperfect performance of the Tiberian tradi-
tion that are reflected by some of the manuscripts of Karaite tran-
scriptions form the British Library. This imperfect performance
was attributed to the impact of the phonological system of the
vernacular language of the scribes. In the current paper I shall
expand on the previous study by examining reflections of imper-
fect performance in a wider range of manuscripts from the British
Library. I shall discuss aspects of imperfect performance discern-
ible in the distribution of the vocalisation signs that are used in
the manuscripts. Many of the Karaite transcriptions have Tibe-
rian vocalisation signs. In several manuscripts these correspond
to the distribution of signs in the standard tradition of Tiberian
vocalisation, as it appears in the model Tiberian masoretic codi-
ces. In many manuscripts, however, some of the signs deviate
from this standard distribution. The paper will focus in particular
on (i) deviations in the distribution of vowel signs that reflect
imperfect performance of Tiberian vowel qualities and (ii) devi-
ations in the distribution of shewa and hatef signs that reflect im-
perfect performance of Tiberian syllable structure. In such man-
uscripts these types of deviation in the use of signs do not take
place in every case and a certain proportion of the marking of
signs corresponds to the standard Tiberian usage.

The corpus that has been used for this study includes the
following manuscripts (BL = British Library):

BL Or 2539 MS A, fols 56-114
BL Or 2549 MS A, fols 1-140
BL Or 2549 MS B, fols 141-308
BL Or 2551 MS A, fols 1-30

BL Or 2551 MS B fols 31-101

206 Khan

BL Or 2552 MS A, fols 1-89
BL Or 2556
BL Or 2559

2.0. Vowel Quality

The Tiberian vowel signs reflect in principle distinctions in qual-
ity (Khan 2020b, I:244-45). Deviations from the standard distri-
bution of the signs could, in principle, reflect either the applica-
tion of the Tiberian signs to represent a different pronunciation
tradition or an inability to distinguish correctly the qualities of
the Tiberian vowels. It is the latter explanation that is the most
satisfactory for the majority of the cases of deviation in distribu-
tion of the vocalisation signs in the corpus of manuscripts studied
in this paper.

The deviations that are found in the manuscripts have been
classified into the following categories:

(1) patah for segol (but not vice versa)

(2) patah-segol interchange

(3) patah-segol interchange, marginal patah-qames inter-
change

(4) patah-segol interchange, patah-qames interchange

2.1. Patah for segol (but not vice versa)
BL Or 2559 fols 1-53
il (BL Or 2559, fol. 5v, 4) || L' wai ‘corpse’ lit. ‘soul’ (Lev. 22.4)

'L = Codex Leningradensis, which is the basis of BHS (Biblia Hebraica
Stuttgartensia). Biblical citations are from BHS unless otherwise indicated.

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 207

all |) (BL Or 2559, fol. 6v, 8) || L ops ‘and you (mpt)’ (Gen. 9.7)

BL Or 2549 MS A fols 140-41
SLally (BL Or 2549, fol. 58r, 6) || L naga’: ‘and for the
beast of’ (Jer. 7.33)

In this manuscript hatef patah occurs in place of hatef segol:

transgress’ (Jer. 2.20)
Ll (BL Or 2549, fol. 22r, 8) || L :nwps I will (not) make’

(Jer. 4.27)

\ga‘Y| (BL Or 2549, fol. 72r, 14) || L xiaby ‘the gods’ (Jer. 10.11)

T7> OTT:

BL Or 2551 MS A, fols 1-30
ck ce! (BL Or 2551 MS A, fol. 21r, 12) || L :qa"nvax ‘4 will
trust in yow’ (Ps. 55.24)

2.2. Patah-segol Interchange

BL Or 2552 MS A, fols 1-89

2.2.1. Patah for segol
ylang (BL Or 2552 MS A, fol. 12r, 11) || L :nan ‘and they
were achat (Job 6.20)

lees (BL Or 2552 MS A, fol. 52r, 8) || Lam ‘he prays’ (Job 33.26)
Jee nn

208 Khan

e-G>lJls (BL Or 2552 MS A, fol. 84v, 11) || L iapnvna
‘will you play with him?’ (Job 40.29)

In this manuscript hatef patah occurs in place of hatef segol:

Me erp

we otree

I do?’ (Job 31.14)

divide him?’ (Job 40.30)
2.2.2. Segol for patah
algsslee (BL Or 2552 MS A, fol. 56r, 9) || L optayn ‘their

works’ (Job 34.25)

2.3. Patah-segol, patah-qames (Marginal) Interchange
2.3.1. Patah for segol
BL Or 2549 MS B fols 141-308
petal 4 (BL Or 2549 MS B, fol. 306r, 8) || L bw7n31 ‘and
in the new moons’ (Ezek. 45.17)
BL Or 2551 MS B fols 31-101
asters > (BL Or 2551 MS B, fol. 41r, 4) || L pS-ama

‘make wide your mouth!’ (Ps. 81.11)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 209

geal (BL Or 2551 MS B, fol. 62r, 14) || L tinx3 ‘they are trust-
worthy’ (Ps. 93.5)
BL Or 2556

a |, 5 (BL Or 2556, fol. 4r, 9) || L yytxa ‘by force’ (Ezra 4.23)

Oj UL (sic with two lams) (BL Or 2556, fol. 16r, 13) || L
TiaIDRwr ‘requires of you’ (Ezra 7.21)

In this manuscript hatef patah occurs in place of hatef segol:

an (BL Or 2556, fol. 15v, 12) || L7a>x ‘your God’ (Ezra 7.19)

pdlselss (BL Or 2556, fol. 69v, 12) || L wtaym ‘and we
Binet (Neh, 10.33)
glde>Le (BL Or 2556, fol. 84r, 12) || L avon ‘made sin (cPLy’
(Neh. 13.26)
dol+-Lusl (BL Or 2556, fol. 112r, 1) || L ton-nwpsx ‘I will
deal loyally’ (1 Chron. 19.2)

2.3.2. Segol for patah

BL Or 2549 MS B fols 141-308

Toe PS

‘gleaming metal’ (Ezek. 8.2)

In this manuscript hatef segol occurs in place of hatef patah:

210 Khan

'Y g«| (BL Or 2549 MS B, fol. 234v, 3) || L nx ‘sick’ (Ezek. 16.30)

BL Or 2551 MS B fols 31-101
|) 3019 (BL Or 2551 MS B, fol. 57v, 4) || L myiawyy ‘and a
teeube 90.4)
< ) 3s) (BL Or 2551 MS B, fol. 68v, 1) || L mtv) ‘for the help
of are on Ps. 102.14)
jolene (BL Or 2551 MS B, fol. 76r, 11) || L inownn ‘his do-
Sisto (Ps. 103.22)
uP! | =u (BL Or 2551 MS B, fol. 32r, 1) || L -pam ‘he will (not)
pa (Exod. 11.7)

BL Or 2556
& i 54s) (BL Or 2556, fol. 83r, 7) || L nitqwsx (ketiv:
ee ee of Ashdod’ (Neh. 13.23)

2.3.3. Qames for patah (Marginal)

BL Or 2549 MS B fols 141-308
| > (BL Or 2549 MS B, fol. 224v, 16) || L m3" ‘it was [not]
cut off (Ezek. 16.4)
L639 (BL Or 2549 MS B, fol. 159v, 12) || L nat ‘to the

g7-3Ts

land of’ (Ezek. 7.2)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 211

In this manuscript hatef patah occurs very marginally in place of
hatef qames:
petlab 3 (BL Or 2549 MS B, fol. 306r, 8) || L bwina ‘and
in the new moons’ (Ezek. 45.17)
BL Or 2551 MS B fols 31-101
Glug (BL Or 2551 MS B, fol. 58v, 10) || L :annon ‘it will be
hidden’ (Isa. 29.14)
BL Or 2556
cab (BL Or 2556, fol. 6v, 2) || L 733 ‘they are building’ (Ezra

iT Tt

5.4)

2.4. Patah-segol, patah-qames Interchange

BL Or 2539 MS A, fols 56-114

2.4.1. Patah for segol
\4 (BL Or 2539 MS A, fol. 63r, 6) || L -nyy ‘and + object
marker’ (Gen. 21.10)
hs '9 (BL Or 2539 MS A, fol. 63v, 8) || L 72m ‘and she went’

(Gen. 21.14)

Cuil (BL Or 2539 MS A, fol. 64r, 3) || Lv) ‘a bow’ (Gen. 21.16)

212 Khan

2.4.2. Segol for patah

os (BL Or 2539 MS A, fol. 63v, 2) || L swan ‘the boy’ (Gen.

- ays

21.12)
Lo 34 (BL Or 2539 MS A, fol. 77r, 9) || L n332m ‘and they
(FPL) walked’ (Gen. 24.61)
ea Bl (BL Or 2539 MS A, fol. 95r, 4) || L_ :qo-1m3 ‘he gave
- you (Ms)’ (Deut. 8.10)

2.4.3. Patah for games

& 134 (BL Or 2539 MS A, fol. 67v, 9) || Lm know’ (Gen. 22.12)

els! | (BL Or 2539 MS A, fol. 68r, 5) || L 07738 ‘Abraham’

ere gS

(Gen. 22.14)

Lislls (BL Or 2539 MS A, fol. 70r, 8) || L 7x7 ‘the woman’

T =

(Gen. 24.5)

cs\yb> (BL Or 2539 MS A, fol. 84r, 1) || L Wat ‘my words’

“AT:

(Deut. 4.10)

he LaLa. (BL Or 2539 MS A, fol. 85v, 6) || L :cxAwnr ‘the heav-

‘l- ore

ens’ (Deut. 4.19)

2.4.4. Qames for patah

GY (BL Or 2539 MS A, fol. 85v, 4) || L p5n ‘he divided’

TOT

(Deut. 4.19)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 213

wel |)) (BL Or 2539 MS A, fol. 94r, 8) || L o-pans ‘forty’ (Deut.

a ae Ha

8.4)

2.5. Discussion

The deviations from the standard distribution of the Tiberian vo-
calisation signs indicate that the scribes were not copying the
signs directly from model Tiberian Bible codices. They must ei-
ther have been copied from manuscripts with a non-standard dis-
tribution of signs or marked independently by the Karaite scribes
in an attempt to represent an oral reading tradition of the text.
In effect, the cause in both scenarios amounts to the same pro-
cess. If they were copied from other manuscripts with non-stand-
ard Tiberian vocalisation, the non-standard distribution in such
manuscripts would itself have been the result of an attempt to
represent an oral reading tradition. It can be assumed, therefore,
that the phenomenon is the result of the assigning of signs to
represent an oral tradition. This oral tradition can be assumed to
be the Tiberian pronunciation tradition. The deviation in distri-
bution is most easily explained as the result of imperfect learning
and performance of the standard Tiberian tradition rather than
the reflection of a different pronunciation tradition, such as the
Palestinian or Babylonian pronunciation, or an extended type of
Tiberian pronunciation tradition. This is because the vocalisation
and transcription of the manuscripts do not reflect distinctive fea-
tures of these other traditions of pronunciation. These would in-
clude features such as the lack of distinction between segol and
sere, which is a feature of the Palestinian pronunciation (Revell
1970), distinctive Babylonian syllabic structure (Yeivin 1985,

214 Khan

283-398), or the extended use of dagesh to all non-guttural con-
sonants as a marker of syllable onset after a preceding closed syl-
lable, which is characteristic of the extended Tiberian tradition
(Morag 1959; Yeivin 1983; Khan 2017).

The various different typologies of deviation in the distri-
bution of the signs from the standard Tiberian vocalisation that
are presented above in 882.1—4 reflect different degrees of imper-
fect learning and performance of the Tiberian pronunciation tra-
dition. The manuscripts in categories 8§2.1-2 exhibit deviations
only with regard to the patah and segol signs. The manuscripts in
categories §§2.3-4, however, exhibit deviations with regard to
the distribution of patah, segol, and qames. It is important to ob-
serve that there is an implicational hierarchy in the typology of
the categories. If there are deviations with regard to qames, this
implies that there are also deviations with regard to patah and
segol. If there are deviations with regard to patah and segol, how-
ever, this does not imply that there is necessarily deviation with
regard to games.

This hierarchy corresponds to different degrees of imper-
fection in the learning and performance of the Tiberian tradition.
Manuscripts with deviation only in the distribution of patah and
segol reflect a lesser degree than those with deviations also with
regard to games.

It can be safely assumed that the vernacular language of
the scribes was Arabic. The fact that some manuscripts reflect
deviations only with regards to patah and segol, which had the
qualities [a] and [e] in the Tiberian pronunciation, indicates that
the Arabic-speaking scribes had greatest difficulty distinguishing
these qualities. This can be explained by the hypothesis that He-
brew [a] and [e] and their respective long counterparts [a:] and

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 215

[e:] were matched by the scribes with the similar sounding Arabic
phonemes /a/ and /a:/. This is a recognised process when two
languages are in contact. It involves the convergence of phono-
logical systems of the languages, whereby phonetic tokens in one
language are matched with a phoneme in a contact language.”
The Arabic phonemes /a/ and /a:/ would have had a range of
allophones, as in the modern Arabic dialects, that included not
only the quality of [a] and [a:], but also the higher quality of [e]
and [e:], by the process of raising (imdla), and the back quality
[a] by the process of suprasegmental pharyngealisation (tafkhim)
(Barkat-Defradas 2011b; 2011a; Levin 2011). This would have
facilitated the interchange of the qualities of Tiberian patah [a]
and [a:] and Tiberian segol [¢] and [e:]. Due to both of these qual-
ities being matched by the Arabic-speaking scribes with the Ara-
bic prototypes [a] and [a:], the speakers had difficulty distin-
guishing their quality in the reading tradition and so imperfectly
applied the standard Tiberian distribution of the signs.

The fact that the scribes were able to maintain the standard
Tiberian distribution of the games and make the correct morpho-
lexical contrasts with patah could be explained by the assumption
that the games phonetic token [9:] that was heard in the Tiberian
reading was not matched with the /a:/ phoneme of Arabic. This
is likely to have been due to its being sufficiently distinct in qual-
ity from the phonetic tokens of Arabic /a:/ for it to be kept apart.
It is a recognised phenomenon in the research of second language
acquisition that learners can more easily acquire a phoneme that
is not similar to one in the native language than a phoneme that
has phonetic tokens that are similar to those of a phoneme in the

? For more details of the process see Blevins (2017).

216 Khan

native language. When there is a high degree of resemblance be-
tween distinct sounds in the target and native languages, they are
more liable to be wrongly matched.* The scribes of manuscripts
in categories §§2.1-2, therefore, correctly learnt the distribution
of Tiberian games and kept it separate from the vowel system of
their Arabic vernacular.

The scribes of manuscripts in categories §§2.3-4, however,
not only failed correctly to learn the Tiberian distribution of
patah and segol, but also imperfectly learnt the distribution of
qames. The vast majority of cases of Tiberian qames that are in-
correctly vocalised in the manuscripts are long qames, but there
are a few sporadic examples of short games. This imperfect learn-
ing and performance would have come about since the scribes
matched also the games with prototypes in the vowel system of
their vernacular speech. These, again, would have been Arabic
/a/ and /a:/. As remarked, Arabic /a/ and /a:/ were realised with
a range of qualities, including [¢] and [e:], by the raising process
of *imdla, and [a] and [a:], by the backing process of tafkhim. The
backed allophones [a] and [a:] occurred in the environment of
the Arabic emphatic, i.e., pharyngealised, consonants, such as /s/
and /t/. The matching of the Hebrew qames vowel, which had
the quality [9], [9:], with Arabic /a/, /a:/, would have been fa-
cilitated by the existence of the similar sounding, though not
identical, backed allophones [a] and [a:] of Arabic /a/ and /a:/.

In order to explain fully the distribution of vowel signs ex-
hibited in the data presented in §§2.1—4, it must be assumed that

3 See, for example, Eckman and Iverson (2003) and the literature cited
there.

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 217

the scribes had learnt the correct phonetic realisation of the Ti-
berian vowel signs (i.e., patah [a], [a:], segol [¢], [e:], games [9],
[9:]). In fact, it is likely that Tiberian patah had a back realisation
[a] in the environment of emphatic consonants such as tet and
tsade (Khan 2020b, I:248), so the scribes would have learnt that
the patah sign had the range of qualities [a, a:, a, a:]. The scribes
did not, however, identify perfectly the sounds of the signs with
what they heard in the reading tradition.

This assumption is necessary to explain why the segol and
patah signs interchange and the patah and qames signs inter-
change, but segol and qames do not interchange, although all
three vowels have been matched with the Arabic prototypes /a/,
fai/.

The realisation of the qualities of the vowel signs in ques-
tion have the following relative position in the buccal vowel
space:

Figure 1: Segol [e]—patah [a, a]—qames [9] in the buccal vowel space

The quality of segol [e] was articulated adjacent to the
range of patah [a, a]. The quality range of patah [a, a] was adja-
cent to both [e] on one side and [5] on the other. The quality of
[e], however, was not adjacent to [5]. The qualities of the Tibe-
rian vowel signs that the scribe had learnt were confused with
qualities adjacent to them in the reading tradition heard by the
scribe. This can be represented as follows:

218 Khan

Table 1: Vowel adjacency and association

associated sounds in the

vowel sien oral reading tradition
1 patah [a, a] Le], [a, a]
2 segol [e] Le], [a, a]
3 patah [a, a] [e], La, a], [o]
4 games [9] La, a], [3]

In manuscripts in category §2.1 only process 1 is attested.
In category §2.2 processes 1 and 2 are attested. In categories
§§2.3-4 all four processes are attested.

The fact that manuscripts in category 82.1 exhibit only the
marking of patah for segol and not vice versa, i.e., process 1, may
possibly be linked to the relative frequency of patah and segol in
the Tiberian Masoretic Text. Patah occurs considerably more fre-
quently than segol. A count of the tokens of patah and segol in the
whole Tiberian Masoretic Text using BibleWorks reveals the fol-
lowing statistics:

patah sign 65,067
segol sign 21,874

This statistical dominance of patah may have made it easier
to confuse segol for patah than patah for segol. Process 1, there-
fore, would be the most liable to occur. The other processes
would be increasingly liable to occur as the degree of imperfect
learning of the reading tradition increased.

It should be remarked that deviation in vocalisation relat-
ing to qames is only marginal in manuscripts of category §2.3.

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 219

This would reflect, therefore, a lower degree of imperfect learn-
ing than is reflected by manuscripts of category §2.4, in which
patah and qames are frequently confused.

3.0. Shewa and hatef Signs

The deviations in the corpus from the standard Tiberian marking
of shewa and hatef signs are presented in various categories be-
low.

3.1. Shewa for hatef
3.1.1. Shewa for hatef patah on Guttural Consonants
BL Or 2539 MS A

| ne (BL Or 2539 MS A, fol. 711, 2) || L my ‘ten’ (Gen. 24.10)

rm! Lal (BL Or 2539 MS A, fol. 71r, 4) || L ov73 ‘Naharaim’
(Gen. 24.10)

me pool iy (BL Or 2539 MS A, fol. 74r, 8) || L :o am ‘and don-
keys’ (Gen. 24.35)

BL Or 2540

Oky 3 (BL Or 2540, fol. 7r, 7) || L tpary ‘you (FPL) have left’
(Exod. 2.20)

-cp> (BL Or 2540, fol. 13r, 2) || L nn ‘bridegroom of (Exod.

4.25)

220 Khan

BL Or 2547
eb iy! '9 (BL Or 2547 fol. 4v, 13) || L op yaw ‘and you (MPL) will
say’ (Josh. 4.7)
elle (BL Or 2547 fol. 5r, 1) || L miaxa ‘the stones’ (Josh. 4.7)
re 5S 19 (BL Or 2547 fol. 2r, 2) || L mins ‘and the priests’
(Josh. 3.14)
g Play (BL Or 2547 fol. 6r, 6) || L 19901 ‘and they (PL) has-
ease: 4.10)

BL Or 2549

sli jae 9h (BL Or 2549, fol. 2r, 1) || L os18"DN3 ‘utterance of
the Lord’ (Jer. 2.19)

Slaslls (BL Or 2549, fol. 58r, 6) || L naaa‘a ‘and for the
beast of (Jer. 7.33)

(slelies (BL Or 2549, fol. 87r, 6) || L *8x3n4 ‘to those who
jisssive cu (Jer. 23.17)

go> (BL Or 2549, fol. 95v, 3) || L 19N9n ‘you will fight’

(Jer. 32.5)

BL Or 2551 MS B

lsL> (BL Or 2551 MS B, fol. 311, 2) || L 5am ‘frost’ (com-

mentary on Ps. 78.47)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 221

BL Or 2552 MS A
Nhl |4 (BL Or 2552 MS A, fol. 10r, 12) || L 7770x1 ‘and I shall
sais (Job 6.10)
yLs'Y (BL Or 2552 MS A, fol. 18r, 12) || L wx ‘into whose’
(Job 12.6)

elas (BL Or 2552 MS A, fol. 85v, 4) || L037 ‘also?’ (Job 41.1)

Loew (BL Or 2552 MS A, fol. 23v, 6) || L mn'q ‘will he live?’
(Job 14.14)
BL Or 2556

UL (BL Or 2556, fol. 6v, 1) || L xvax ‘we said’ (Ezra 5.4)

T: 473

cols (BL Or 2556, fol. 19v, 5) || L>as1 ‘and P (Ezra 7.28)
21 4819 (BL Or 2556, fol. 4v, 2) || L mm ‘and it was’ (Ezra 4.24)

LEalS (BL Or 2556, fol. 12r, 9) || L x9 ‘the priests’ (Ezra 6.16)

tT: ~T

3.1.2. Shewa for hatef patah on Non-guttural Consonants in L
BL Or 2549 MS B

will eat it (Ezek. 4.9)

222 Khan

les 9 (BL Or 2549 MS B, fol. 177v, 8) || L wNwxn ‘(D was
left Ezek. 9.8)

BL Or 2551 MS A
Sd % (BL Or 2551 MS A, fol. 19r, 12) || L *y-aqpn ‘and
from ms against me’ (Ps. 55.20)
jek 59 (BL Or 2551 MS A, fol. 20r, 10) || L iad-aqp3 ‘and war
was : his heart’ (Ps. 55.22)

BL Or 2551 MS B
s)l (BL Or 2551 Ms B, fol. 63v, 11) || L 1992 ‘bless! (MPL)’
Ps, 100.4)

BL Or 2552 MS A
ee oe (BL Or 2552 MS A, fol. 37v, 13) || L ‘292 ‘they (did
not) ee us’ (Job 31.20)
re L)\8] (BL Or 2552 MS A, fol. 42r, 12) || L s3a7px ‘I will go
sate 55 him’ (Job 31.37)
_jilab y) (BL Or 2552 MS A, fol. 52r, 4) || L wav ‘it will be
fresh’ (Job 33.25)
jll.e (BL Or 2552 MS A, fol. 83v, 4) || L iby ‘his shadow’

(Job 40.22)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 223

BL Or 2556
lbeal | 4 (BL Or 2556, fol. 1r, 6) || L oa ‘and to teach them’
(Dan. 1.4)
S| |4 (BL Or 2556, fol. 22r, 7) || L n?2wyi ‘and I weighed
si fies 8.26)
g> lel (BL Or 2556, fol. 75v, 10) || L 127371 ‘And they (m)
piace (Neh. 11.2)
gle (BL Or 2556, fol. 81r, 8) || L455 ‘began to be dark (MPL)’

(Neh. 13.9)

3.1.3. Shewa for hatef segol on Guttural Consonants

BL Or 2539
ne oa (BL Or 2539 MS A, fol. 66v, 6) || L :o*75x7 ‘God’ (Gen.
22.3)

BL Or 2547
(Sg (BL Or 2547 fol. 15r, 2) || L :qaxq ‘the Amorites’
(Josh. 13.4)
ges) (BL Or 2547 fol. 18v, 15) || L 12°78 ‘they outlived’
(iudg. 2.7)

e§5! (BL Or 2547 fol. 29v, 16) || L inty ‘Edom’ (Judg. 11.17)

224 Khan

BL Or 2552 MS A

342) (BL Or 2552 MS A, fol. 43v, 5) || L xix ‘Elihu’ (Job 32.2)

33.26)
(sa 56 (BL Or 2552 MS A, fol. 52v, 7) || L siya ‘I have

perverted’ (Job 33.27)

BL Or 2556

ene &

Sey (BL Or 2556, fol. 3r, 3) || L x19 ‘will be’ (Dan. 2.28)

La'Y\ (BL Or 2556, fol. 4v, 2) || L xadx ‘the God’ (Ezra 4.24)

TO OT:

cyl5ly (BL Or 2556, fol. 6v, 13) || L rx1 ‘and then’ (Ezra 5.5)

3.1.4. Shewa for hatef segol on Non-guttural Consonants in L

3.1.5. Shewa for hatef games on Guttural Consonants

BL Or 2552 MS A

Al (BL Or 2552 MS A, fol. 40r, 4) || L-2n¥ ‘my tent’ Job 31.31)

AST

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions

225

3.1.6. Shewa for hatef games on Non-guttural Consonants in L

BL Or 2556

revit han (BL Or 2556, fol. 122v, 1) || L wp ‘the dedicated
gifts’ (1 Chron. 28.12)

oF oe tas (BL Or 2556, fol. 4r, 7) || L nintvo7p ‘before Re-
hum’ (Ezra 4.23)

he8Y (BL Or 2556, fol. 11r, 12) || L bap ‘according to’ (Ezra
6.13)

e3-JS (BL Or 2556, fol. 15r, 11) || L 5ap-43 ‘in accordance
wie’ (Ezra 7.17)

lb (BL Or 2556, fol. 14r, 13) || L D4p ‘before’ (Ezra 7.14)

BL Or 2559

| Lia (BL Or 2559, fol. 3v, 5) || L bwtpa ‘the sacred do-
nations’ (Lev. 22.3)
5\a (BL Or 2559, fol. 5r, 12) || L bps ‘of the sacred

donations’ (Lev. 22.4)

3.2. Shewa for Vowel in Unstressed Closed Syllables

Shewa occurs for patah in closed unstressed syllables in L:

226 Khan

BL Or 2539 MS A
Ge ) (BL Or 2539 MS A, fol. 73v, 8) || L nman ‘and he
opened’ (Gen. 24.32)
leox3'Y y (BL Or 2539 MS A, fol. 74v, 4) || L pnp) ‘and you will
ie (Gen, 24.38)
>? §4 (BL Or 2539 MS A, fol. 74v, 6) || L 7752 ‘in the morning’
(Ge 24.54)

BL Or 2540

\| ® (BL Or 2540, fol. 8r, 4) || L nya ‘the sight’ (Exod. 3.3)
an (BL Or 2540, fol. 8r, 2) || L 7857 ‘the angel of (Exod. 3.2)

BL Or 2551 MS B
a , (BL Or 2551 MS B, fol. 32r, 6) || L 7% ‘and he smote’ (Ps.
78.51)
Le L..8 9 (BL Or 2551 MS B, fol. 39r, 13) || L n22anm ‘and
the eingaaae (Isa. 60.12)
he ¥ yee (BL Or 2551 MS B, fol. 80v, 4) || L :ova-sy» ‘the
waters stood’ (Ps. 104.6)

BL Or 2556
omen (BL Or 2556, fol. 116v, 8) || L atmo ‘for nails’ (1

Chron. 22.3)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 227

3.3. Hatef for shewa in L
Hatef patah occurs for shewa on non-guttural consonants in L:

BL Or 2539 MS A

| Lis (BL Or 2539 MS A, fol. 71v, 9) || L “wm ‘and the gir!’

Rs SSE

(Gen. 24.16)

‘Y se (BL Or 2539 MS A, fol. 93r, 2) || L743 ‘great’ (Deut. 7.23)

BL Or 2549 MS B
Pit _S (BL Or 2549 MS B, fol. 174v, 10) || L 3X)" ‘be-
aes they filled’ (Ezek. 8.17)
els j4 (BL Or 2549 MS B, fol. 157r, 1) || L ngzina ‘in the

midst of you’ (Ezek. 6.7)

3.4. Hatef for Vowel in Unstressed Closed Syllables
3.4.1. Hatef patah for patah in Unstressed Closed Syllable
BL Or 2539 MS A
> § (BL Or 2539 MS A, fol. 74v, 6) || L712 ‘in the morning’
feat 24.54)
leox3'Y 4 (BL Or 2539 MS A, fol. 74v, 8) || L nnpt ‘you (ms)
will Gis (Gen. 24.40)
vel ae (BL Or 2539 MS A, fol. 78v, 5) || L Dwxy57 ‘the

concubines’ (Gen. 25.6)

228 Khan

pleusle) (BL Or 2539 MS A, fol. 81r, 5) || L ndzvn ‘because
of you (MPL)’ (Deut. 3.26)

BL Or 2551 MS B
opine (BL Or 2551 MS B, fol. 83v, 1) || L 32Wyr ‘let them
[not] reer me’ (Ps. 119.122)
33) 4 (BL Or 2551 MS B, fol. 34r, 12) || L imam ‘and his
ae (s. 78.61)

a (BL Or 2551 MSB, fol. 38v, 10) || Lav4 ‘mocking’ (Ps. 79.4)

L uel (BL Or 2551 MS B, fol. 46r, 1) || L you ‘your anger’
Ps. 85.5)

3.4.2. Hatef segol for segol in Unstressed Closed Syllable

BL Or 2551 MS B

$l (BL Or 2551 MS B, fol. 33r, 13) || L try ‘high’ (Ps. 78.56)

es J) (BL Or 2551 MS B, fol. 47r, 2) || L tay-bx ‘to his peo-

ple’ (Ps. 85.9)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 229

3.5. Hatef for Vowel in Stressed Closed Syllables

3.5.1. Hatef patah for patah in Stressed Closed Syllables

BL Or 2551 MS B
(ala! (BL Or 2551 MS B, fol. 82v, 4) || L ‘many ‘I love’ (Ps.
119.119)

3.5.2. Hatef games for games in Stressed Closed Syllable

BL Or 2551 MS B
(ov) (BL Or 2551 MS B, fol. 90r, 2) || L »many ‘I love’ (Ps.

*: 0 TIT

119.163)

3.6. Vowel for shewa
Patah is marked in place of shewa in a number of manuscripts:
BL Or 2539 MS A
re | 2x9 (BL Or 2539 MS A, fol. 65v, 3) || L ama" ‘and they
(MPL) cut off (Gen. 21.27)
ile _y» (BL Or 2539 MS A, fol. 66r, 3) || L ikayWw ‘chief of
his army’ (Gen. 21.32)
oeald (BL Or 2539 MS A, fol. 66r, 4) || L :o’nw4a ‘Philistines’

(Gen. 21.32)

230 Khan

BL Or 2540
Lo ySins (BL Or 2540, fol. 4r, 7) || L nizg9n ‘supplies’ (Exod.
1.11) |
ge _pial 9 (BL Or 2540, fol. 3v, 7) || L ayqwe1 ‘and they (pL)
evuaced (Exod. 1.7)

BL Or 2551 MS B
3 y-4-Y (BL Or 2551 MS B, fol. 35v, 10) || L 32 77-N} ‘they
will - tread’ (1 Sam. 5.5)
mu jie) (BL Or 2551 MS B, fol. 55v, 1) || L own ‘the
singers’ (commentary on Ps. 87.7)

BL Or 2559
G 539 (BL Or 2559, fol. 3v, 12) || L aman ‘and she shall be

cut off? (Lev. 22.3)

3.7. Vowel for hatef

3.7.1. Patah for hatef patah on Guttural Consonants

BL Or 2539 MS A

(+ ULI (BL Or 2539 MS A, fol. 63v, 2) || L qhnax- ‘your hand-

maid’ (Gen. 21.12)

Ls; (BL Or 2539 MS A, fol. 63v, 5) || L yun ‘your seed’

CG, ae i!

(Gen. 21.13)

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 231

cS 5>leeS (BL Or 2539 MS A, fol. 64r, 3) || L 1099 ‘like the
shots of (Gea 21.16)

BL Or 2549
S j>kig (BL Or 2549, fol. 47v, 8) || L ninmwn> ‘to wor-
ship’ (Jer. 72)

BL Or 2551 MSA

co! (BL Or 2551 MS A, fol. 9v, 10) || L I-31 ‘but P’ (Ps. 52.10)

olsl) (BL Or 2551 MS A, fol. 9v, 10) || L 1277, ‘green’ (Ps. 52.10)

T- -

| 8) (BL Or 2551 MS A, fol. 10r, 7) || L 721 ‘and I will wait’ (Ps.

52.11)

spare

BL Or 2551 MS B
U&! (BL Or 2551 MS B, fol. 37r, 12) || L ¥pax ‘I will seek’
(Ezek. 34.16)
lL > (BL Or 2551 MS B, fol. 58v, 10) || L 727 ‘its wise

5 et eel

men’ (Isa. 29.14)

119.124)

232 Khan

ged (BL Or 2551 MS B, fol. 39v, 3) || L 132°w7 ‘restore us!
(ms)’ (Ps. 80.20)
BL Or 2559

Ls] (BL Or 2559, fol. 6v, 2) || L 1wx ‘which’ (Lev. 22.5)

3.7.2. Patah for hatef patah on Non-guttural Consonants in L

BL Or 2539 MS A
g> |48 9 (BL Or 2539 MS A, fol. 68v, 4) || L 197an7m ‘and they
will bless themselves (MPL)’ (Gen. 22.18)

BL Or 2551 MS B
sey (BL Or 2551 MS B, fol. 99r, 1) || L 17734 ‘and bless!
(MPL)’ (Ps. 134.2)

3.7.3. Segol for hatef segol on Guttural Consonants

BL Or 2539 MS A
jl+\G (BL Or 2539 MS A, fol. 68r, 3) || L mx3 ‘it was caught’
(aa: 22.13)

BL Or 2540

3
we 3) (BL Or 2540, fol. 8v, 6) || L byidy ‘God’ (Exod. 3.14)

ott

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 233

BL Or 2549 MS B

CS) 5¢| (BL Or 2549 MS B, fol. 238v, 3) || L :hx ‘Amorite’

(Ezek. 16.45)

BL Or 2551 MS A

ost J) (BL Or 2551 MS A, fol. 9v, 3) || L oraox ‘God’ (Ps. 52.9)

G 5a..8 (BL Or 2551 MS A, fol. 12v, 3) || L nnaq ‘you put to
shame’ (Ps. 53.6)

7 #) (BL Or 2551 MS A, fol. 17v, 1) || L wig ‘man’ (Ps. 55.14)

(Ps. 56.1)

BL Or 2551 MS B
U squid (BL Or 2551 MS B, fol. 45v, 2) || L niaw7 ‘you caused
to sabia (Ps. 85.4)

3.7.4. Qames for hatef games on Guttural Consonants

BL Or 2540

i= (BL Or 2540, fol. 8v, 6) || L *3p ‘affliction’ (Exod. 3.7)

BL Or 2551 MS A
he | ale 5 (BL Or 2551 MS A, fol. 19r, 8) || L o7¥1,‘and noon’

(Ps. 55.18)

234 Khan

BL Or 2551 MS B
pili (BL Or 2551 MS B, fol. 81v, 7) || L o'xay ‘foliage’ (Ps.
104.12)
3.7.5. Qames for hatef games on Non-guttural Consonants in L
BL Or 2556
pee (BL Or 2556, fol. 83r, 7) || L nizay (ketiv: nvnny)

‘women of Ammon’ (Neh. 13.23)

3.8. Discussion

In the Tiberian pronunciation tradition, a vocalic shewa in prin-
ciple represents a short vowel in an open syllable (CV).* Its qual-
ity was by default the same as that of the patah vowel sign, i.e.,
the maximally low vowel [a], e.g.,

neon [t"ayas'se:] ‘you (Ms) cover’ (Job 21.26)

oy 271 [madabba'ri:im] ‘speaking’ (MPL) (Est. 2.14)

When vocalic shewa occurs before a guttural consonant or
the letter yod, it was realised with different qualities through as-
similatory processes. Before a guttural (i.e., &, 7, n, 9) it was re-
alised as a short vowel with the quality of the vowel on the gut-

tural, e.g.,
ya wa [beferk"a'yo:] ‘by your evaluation’ (Lev. 5.15)
mm [vaha:'ja:] ‘and it became’ (Gen. 2.10)
782 [be'?e:er] ‘well’

* For further details concerning shewa and hatef vowels in the Tiberian
pronunciation tradition, see Khan (2020b, I:305-47).

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 235

tix [mo'?0:06] ‘very’

Before yod, it was realised as a short vowel with the quality

of short hiregq [i], e.g.,

ova [bi'jo:om] ‘on the day’ (Gen. 2.17)

Sew [lijisr*o:'?e:el] ‘to Israel (Gen. 46.2)

The shewa sign is combined with some of the basic vowel
signs to form the so-called hatef signs. In such signs the vocalic
reading of the shewa as well as its quality are made explicit. The
vocalic shewa and the hatef vowels were quantitatively equiva-
lent. In all cases they form short open syllables (CV).

In the Tiberian pronunciation the CV of a vocalic shewa or
a hatef vowel cannot stand alone, but is prosodically dependent
on the following syllable, which must be bimoraic (CVV or CVC).
The CV syllable is bound with the following syllable in a single
metrical foot. It is a metrically weak syllable and the following
bimoraic syllable is the strong syllable of the foot. This can be
represented thus: (. *), where the brackets enclose the syllables
of the foot, the star * represents the strong prominent syllable,
and the dot the weak syllable. On a prosodic level, therefore, the
phonetic realisation of a word such as 750n [t"isp"a'ru:] would
consist of three syllables parsed into two feet:

[(tis.) (p*a.'Ru:)]

(*) C)

This dependent prosodic status of vocalic shewa and hatef
vowels is associated with the fact that they have the status of
epenthetic vowels that break up consonant clusters at syllable
onset. On an underlying phonological level, a word such as 750m

[tisp"a'Ru:] would have the form /tispru:/, with the shewa [a] as

236 Khan

an epenthetic that breaks the onset cluster /pr/ on the phonetic
level. The fact that vocalic shewa is zero on the phonological level
appears to be the reason why the Masoretes marked it with the
same sign as they used to mark silent shewa. The hatef signs ap-
pear to be later developments of the notation system that made
the reading of shewa as vocalic explicit in certain contexts.

Some of the deviations from the standard Tiberian vocali-
sation with regard to shewa and hatef vowels that are presented
above from the Karaite manuscripts may be regarded as reflect-
ing a more primitive stage of the development of the Tiberian
vocalisation system. This may apply to the marking of shewa in-
stead of a hatef sign on guttural consonants (83.1), in which the
reading of a shewa on a guttural was not marked explicitly as
vocalic by the addition of a vowel sign next to the shewa sign.
This phenomenon is found in many Hebrew manuscripts in He-
brew script with Non-Standard Tiberian vocalisation (Khan
2020b, I:340). This may also apply to the marking of shewa where
L has a hatef sign on a non-guttural consonant (§§83.1.2, 3.1.4,
§3.1.6). The model masoretic codices are not consistent in the
marking of hatef in this context and some have shewa where L has
a hatef (Khan 2020b, 1:343-46).

The majority of the deviations, however, can be explained
as being the result of a reanalysis of the syllable structure in the
Tiberian pronunciation. This reanalysis resulted in shewa and
hatef being interpreted as short vowels on the phonological level
rather than phonetic epenthetic vowels. They were, therefore,
equivalent to short vowels in closed CVC syllables. This arose

since the monomoraic syllable CV with shewa or hatef vowels

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 237

came to be analysed as a legitimate syllable on the phonological
level. As is the case with the phonological reanalysis of the qual-
ity of vowels, the reanalysis of CV as a legitimate phonological
syllable is likely to have been induced by convergence with the
phonological system of Arabic, which was the vernacular of the
scribes. In Classical Arabic and also in the modern eastern Arabic
dialects, such as those of Egypt and the Levant, CV is a legitimate
syllable at the phonological level, whether stressed or unstressed,
e.g. Modern Cairene Arabic: kd.tab ‘he wrote’, ka.tdbt ‘I/you (ms)
wrote’ (Mitchell 1962, 26; Watson 2007, 56-58). A word such as
m0n [t"asapp"a'ru:] in the Tiberian pronunciation would have
the phonological syllable structure /tsappru:/. If, however,
[ttasapp*a'ru:] were parsed according to Arabic syllabic princi-
ples, the CV syllables would be analysed as phonological syllables
rather than the result of phonetic epenthesis, thus /tasapparu:/.
As a result, the /a/ in the open CV syllables /ta/ and /pa/ would
be interpreted as having the same phonological status as the /a/
in the closed syllable /sap/. It would follow from this reanalysis
that a shewa sign and a patah sign in a closed syllable represented
vowels that were equivalent and this facilitated the interchange
of the signs. The same would apply to hatef signs, which, after
this syllabic reanalysis according to Arabic principles, would
come to be interpreted as representing vowels that were equiva-
lent to the vowel represented by a vowel sign of the same quality
that is used to represent a short vowel in unstressed closed sylla-
bles, e.g., in a word such as O77 ‘you (MPL) spoke’ (Gen. 43.27).
According to this Arabic type of parsing of syllable structure, the

notational distinction between shewa, hatef, and full vowel signs

238 Khan

lost its original function of distinguishing between phonological
vowels and phonetic epenthetics, and so the signs were freely in-
terchanged in open CV and closed CVC syllables.

All cases of shewa marked in closed syllables in the corpus
are in unstressed closed syllables, in which the vowel would have
been short. The vast majority of cases of hatef signs in closed syl-
lables are likewise in unstressed syllables. There are only two

cases in stressed syllables, viz.,

BL Or 2551 MS B
(us| (BL Or 2551 MS B, fol. 82v, 4) || L mans ‘T loved’
(Ps. 119.119)

BL Or 2551 MS B
(v4! (BL Or 2551 MS B, fol. 90r, 2) || L »maqx ‘T loved’

(Ps. 119.163)

The practice of marking shewa and hatef signs in closed un-
stressed syllables is sporadically found even in some of the Stand-
ard Tiberian Masoretic codices,° e.g.

nav na ‘on the magicians’ (L Exod. 9.11)

mata ‘the evening’ (L Exod. 30.8)

iptm ‘they are strong’ (L 2 Sam. 10.11)

72” ‘he brings trouble on you’ (L Josh. 7.25)

4337 ‘and we will kill him’ (L Judg. 16.2)

72! (BL Or. 4445) || L 72! ‘to Molech’ (Lev. 20.3)

° Yeivin (1968, 18), Dotan (1985).

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 239

This practice in the vocalisation of the model codices may
also have been facilitated by contact with Arabic syllable struc-

ture, as described above.

4.0. Concluding Remarks

In this paper I have presented various examples of the use of Ti-
berian vocalisation signs in the Karaite transcriptions of the He-
brew Bible into Arabic script. The focus in the paper has been on
cases of vocalisation signs in the manuscripts that deviate from
the distribution of the signs that are found in the Standard Tibe-
rian Masoretic tradition. These deviations relate to the distribu-
tion of signs representing different vowel qualities and to the dis-
tribution of shewa and hatef signs. In both sets of cases, it was
argued that the deviations can be explained by the hypothesis
that the Hebrew of the scribes had undergone a convergence with
the phonological structure of their Arabic vernacular. In the case
of vowel qualities, this convergence would have resulted in diffi-
culties in distinguishing between some of the Tiberian vowel
qualities. In the case of shewa and hatef vowels, the convergence
resulted in a reanalysis of epenthetic CV syllables of shewa and
hatef as phonological syllables. It followed that the distinction
between shewa and hatef signs in open CV syllables and vowel
signs in CVC syllables became redundant and the signs, therefore,

were interchanged.

References

Barkat-Defradas, Melissa. 2011a. ‘Vowel Backing’. In Encyclo-
pedia of Arabic Language and Linguistics, edited by Lutz

240 Khan

Edzard and Rudolf de Jong. online: Brill.

http://dx.doi.org.ezp.lib.cam.ac.uk/10.1163/1570-

6699_eall_EALL_COM_0371.

. 201 1b. ‘Vowel Raising’. In Encyclopedia of Arabic Language and
Linguistics, edited by Lutz Edzard and Rudolf de Jong. online:
Brill. http://dx.doi.org.ezp.lib.cam.ac.uk/10.1163/1570-
6699 eall EALL SIM_0140.

Blevins, Juliette. 2017. ‘Areal Sound Patterns: From Perceptual
Magnets to Stone Soup’. In The Cambridge Handbook of Ar-
eal Linguistics, edited by Raymond Hickey, 88-121. Cam-

bridge: Cambridge University Press.

Dotan, Aron. 1985. ‘Pathé Hatfin: A Study in the Evolution of the
Tiberian Vocalization’. In Avraham Even-Shoshan Volume,
edited by Ben-Zion Luria, 157-65. Jerusalem: Kiryath Se-
pher. [Hebrew]

Eckman, Fred R., Abdullah Elreyes, and Gregory K. Iverson.
2003. ‘Some Principles of Second Language Phonology’.
Second Language Research 19/3: 169-208.

Harviainen, Tapani. 1993. ‘Karaite Arabic Transcriptions of He-
brew in the Saltykov-Shchedrin Public Library in St. Peters-
burg.’ In Estudios Masorericos: En Memoria de Harry M. Or-
linsky, edited by Emilia Fernandez Tejero and Maria Teresa
Ortega Monasterio, 63-72. Textos y Estudios «Cardenal Cis-
neros» de La Biblia Poliglota Matritense 55. Madrid: Con-
sejo Superior de Investigaciones Cientificos.

Khan, Geoffrey. 1990. Karaite Bible Manuscripts from the Cairo Ge-
nizah. Cambridge University Library Genizah Series 9.

Cambridge: Cambridge University Press.

Hebrew Vocalisation in Karaite Arabic-Script Transcriptions 241

. 1992. ‘The Medieval Karaite Transcriptions of Hebrew in
Arabic Script’. Israel Oriental Studies 12: 157-76.

. 1993. ‘On the Question of Script in Medieval Karaite
Manuscripts: New Evidence from the Genizah’. Bulletin of
the John Rylands University Library of Manchester 75: 133-
41.

. 2017. ‘The Background of the So-Called Extended Tiber-
ian Vocalization of Hebrew’. Journal of Near Eastern Studies
76/2: 265-73.

. 2020a. ‘Some Features of the Imperfect Oral Performance

of the Tiberian Reading Tradition of Biblical Hebrew in the
Middle Ages’. In Studies in Semitic Vocalisation and Reading
Traditions, edited by Aaron Hornkohl and Geoffrey Khan,
549-92. Cambridge Semitic Languages and Cultures 3.
Cambridge: University of Cambridge and Open Book Pub-
lishers.

. 2020b. The Tiberian Pronunciation Tradition of Biblical He-
brew: Including a Critical Edition and English Translation of

the Sections on Consonants and Vowels in the Masoretic Trea-
tise Hidayat al-Qari’ ‘Guide for the Reader’. 2 vols. Cambridge
Semitic Languages and Cultures 1. Cambridge: University
of Cambridge and Open Book _ Publishers.
doi.org/10.11647/OBP.0163.

Levin, Aryeh. 2011. “Imala’. In Encyclopedia of Arabic Language and
Linguistics, edited by Lutz Edzard and Rudolf de Jong. online:
Brill. http://dx.doi.org.ezp.lib.cam.ac.uk/10.1163/1570-
6699_eall_ EALL_ SIM_vol2_0022.

242 Khan

Mitchell, Terry F. 1962. Colloquial Arabic: The Living Language of
Egypt. London: Teach Yourself Books.

Morag, Shelomo. 1959. ‘The Vocalization of Codex Reuchlinia-
nus: Is the Pre-Masoretic Bible Pre-Masoretic?’ Journal of
Semitic Studies 4: 216-37.

Revell, E. John. 1970. ‘Studies in the Palestinian Vocalization of
Hebrew’. In Essays on the Ancient Semitic World, edited by
John. W. Wevers and Donald. B. Redford, 59-100. Toronto:
Toronto University Press.

Tirosh-Becker, Ofra. 2011. Rabbinic Excerpts in Medieval Karaite
Literature. 2 vols. Jerusalem: Bialik Institute and the He-
brew University. [Hebrew]

Watson, Janet C. E. 2007. The Phonology and Morphology of Ara-
bic. Oxford: Oxford University Press.

Yeivin, Israel. 1968. Aleppo Codex. Jerusalem: Magnes Press and
The Hebrew University of Jerusalem. [Hebrew]

. 1983. ‘Mashmaut siman ha-dagesh ba-niqud ha-tavrani ha-

murhav’. In Hebrew Language Studies Presented to Professor
Zeev Ben-Hayyim, edited by Moshe Bar-Asher, Aron Dotan,
David Tene, and Gad Ben-Sarfatti, 293-307. Jerusalem:
Magnes.

. 1985. The Hebrew Language Tradition as Reflected in the

Babylonian Vocalization. Jerusalem: The Academy of the

Hebrew Language. [Hebrew]

DISSONANCE BETWEEN MASORETIC
VOCALISATION AND CANTILLATION IN
BIBLICAL VERSE DIVISION

Yochanan Breuer

1.0. Introduction

The Masora of the Hebrew Bible rests on three pillars: consonants
or written form (ketiv), vocalisation (niqqud), and cantillation
(te‘amim) which combine to produce the Masoretic Text.

These three facets are not separate, but inextricably inter-
connected. It is impossible to vocalise a biblical verse without
first clarifying its written form, and it is impossible to cantillate
a verse before its vocalisation has been set.

The cantillation depends on the vocalisation, by way of ex-
ample, in the rules for exchanging of disjunctives and in deter-
mining some of the conjunctive cantillations. These all depend
on word length —and the precise length of a word can be deter-
mined only in accordance with its precise vocalisation. Equally,
vocalisation depends on cantillation: for example, in the rules for
the pronunciation of the begadkefat consonants at the beginning
of a word after an open syllable, for which purpose it is important
to know whether the preceding word has a disjunctive or a con-
junctive cantillation. Similarly, pausal forms depend on the main

disjunctive accents.

244 Breuer

When the Masoretes embarked on the task of determining
the precise format of the biblical text, they were, accordingly,
required to establish this format not in a single aspect, but in all
these three aspects together. In each verse, their task was not
only to fix the spelling, vocalisation, and cantillation, but also to
ask themselves whether these three foundations were consistent.
Although the agreement between the three foundations is indeed
firm in most cases, there are instances when each of these foun-
dations heads in a different direction.

The most prominent example of dissonance between the
different foundations is the phenomenon of the distinction be-
tween the written form (ketiv) and the form that is read out (qere).
Dissonance can also be found between the ketiv and the cantilla-
tion, albeit only in rare instances (M. Breuer 1981). In this article
I intend to show this same phenomenon, but this time regarding
the connection between the vocalisation and the cantillation. A
careful observer of the meaning of the biblical text as indicated
by the vocalisation will find that this is not always consistent
with the meaning dictated by the cantillation.

For each verse I have attempted to present the disagree-
ments among the biblical commentators regarding its interpreta-
tion. This will add to the objectivity of my examination: the two
interpretations I propose in each case have not been invented
merely to resolve the dissonance between the cantillation and the
vocalisation. These differences were present in the exegetical lit-
erature. This means that such cases of dissonance are not to be

seen as results of artificial masoretic interpretation. Instead, it is

Dissonance between Masoretic Accentuation and Cantillation 245

an expression of a genuine problem in the interpretation of the
verse.’

What follows is not an exhaustive list of verses in the He-
brew Bible that display dissonance between vocalisation and can-
tillation. Instead, the verses below exemplify four different areas
in which dissonance occurs:

Absolute versus construct: word groupings that were re-
garded by the cantillators as construct phrases, but whose words
were separated by the vocalisers, or vice versa;

Definiteness: words that were considered definite by the vo-
calisers, but indefinite by the cantillators;

Pausal versus non-pausal: words that are vocalised as pausal
forms, but cantillated with weak disjunctive (or conjunctive) ac-
cents, and words vocalised as non-pausal, but cantillated with
strong disjunctives;

One word versus two words: words perceived as a single word

by the vocalisers, but as two by the cantillators.
2.0. Absolute versus Construct

2.1. Construct in the Vocalisation, Absolute in the

Cantillation

(1) ste ya nowy odaNT wong hnan onhomns yas oon ney
‘Opk MWY AAT Nivawn nop
‘With the work of an engraver in stone, like the engravings
of a signet, shalt thou engrave the two stones, according to

' The version of the biblical text is according to M. Breuer (1989). The
Aramaic Targum versions quoted are according to Sperber 1959-1973.

246 Breuer

the names of the children of Israel; thou shalt make them
to be enclosed in settings of gold.’ (Exod. 28.11)?

The vocalisation of wan shows that we have here a construct:
jaX-win, i.e. “an engraver in stone” or a “stone-engraver.” In
other words, the onyx stones should be made by a stone-en-
graver. The division of the vocalised text is therefore:

jax wan nwyn
Rashi interpreted the verse as follows:

apnd> gin pat at wan .orasx Sw jax Awyn - jax won Awyn
Sv won ip n01 oy won a1, 1a1D2 nna Tip gin Jas ,panxdw
pminar opiat 7x 52 ,tEyn Srna won 72) .oy

The work of a master of stones. This win is affixed to the
following word, and accordingly its final vowel is a patah.
Similarly, we have 1p nvi oxy win (Isa 44:13), “an en-
graver of wood”; and txyn Sra win “an engraver of iron”
(Isa. 44.12); all these are in the construct and vocalised
with a patah.

Rabbi Shmuel Ben Meir offered the same commentary.

Yet the cantillation divides the verse differently:
jax dan vwyn

The meaning here is that the onyx stones should be fashioned by
an engraver, and that they should be made of stone. Here, then,
Tax ‘stone’ is a complement of wy ‘work’. This same meaning is
conveyed by the Aramaic translations: xn”>370 ,j~17v 718 Tay ‘the

work of an artisan the pearls will be’ (Ps.-Jonathan); wsav

? The English translations are taken from the online edition of Mechon
Mamre (https://www.mechon-mamre.org/p/pt/pt0.htm).

Dissonance between Masoretic Accentuation and Cantillation 247

holy ara .iar’s ‘the work of an artisan, engraved stones’
(Pesh.).

Luzzatto sensed the contradiction between the vocalisation
and the cantillation in this verse, commenting that the zargqa
should be on nwyn rather than on win (he quotes Rashi on the
verse). Since he interpreted the verse according to the vocalisa-
tion, he failed to recognise that the cantillation reflects a different
interpretation and should not be corrected.

Which interpretation embodies the plain meaning of the
verse? It is difficult to determine. Rashi cites n’xy win ‘carpenter’
and 51a win ‘blacksmith’ in support of the construct form wan
jax, as indicated by the vocalisation. We should add that it is
possible to find constructs that are extremely similar to j28 won
in:

TAS WITT PR WIN) OS DT TTR DNA W779 OPN Nw

“TIT? Mayan Ve

‘and Hiram king of Tyre sent messengers to David, and ce-

dar-trees, and carpenters, and masons—and they built a

house for David.’ (2 Sam. 5.11)

In the account of the making of the Tabernacle itself, we find:

---PR NANA NN? TA NWT
‘and in cutting of stones for setting, and in carving of
wood...’ (Exod. 31.5; 35.33)

Accordingly, it is possible that here, too, we have a construct
form ]3X-win.

However, there is another subject that must be examined
regarding this verse. The verse includes the root nwy (do, make)

and the noun jax (stone). We must ask what the connection is

248 Breuer

between the verb nwy, which appears frequently in the descrip-
tion of the building of the Tabernacle, and the material from
which the tools of the Tabernacle are made. Is the material con-
nected to the verb nwy, as implied here by the cantillation: -nwyn
jax — win, or is it not connected to nwy, in which case the ap-
proach is parallel to that of the vocalisation, which does not con-
nect jax and nwyn. We find divergent practices in this regard.

In some instances, the material is connected to nwy:

- ANT ODD OW ny)

‘And thou shalt make two cherubim with gold...’ (Exod.
25.18)

DOW (ZY NAVAN Mwy)

‘And thou shalt make the altar with acacia-wood...’ (Exod.
27.1)

DOW (YD DTT NS Mwy

‘And thou shalt make the staves with acacia-wood...’ (Exod.
30.5)

DOW "YD NOP] Naa ny wy

‘And he made the altar of incense with acacia-wood...’
(Exod. 37.25)

DOW "YY TIvA Narny wyn

‘And he made the altar of burnt-offering with acacia-
wood...’ (Exod. 38.1)

In these verses, each material is clearly connected with nwy.

Elsewhere, however, we find that although the verb nwy
and the material appear in the same verse, they are not connected
to each other. In these instances, the material is bound to the
name of the item being made in a construct phrase:

Dissonance between Masoretic Accentuation and Cantillation 249

Ao wan Mby
‘And thou shalt make fifty clasps of gold...’ (Exod. 26.6)
230 nyzin) yas) Nan jyaq yyweoy oryy
‘And upon the skirts of it thou shalt make pomegranates of
blue, and of purple, and of scarlet...’ (Exod. 28.33)
T-DNA BA? Aw
‘And thou shalt make them linen breeches...’ (Exod. 28.42)
2 THO ANT AVA WY"
‘And they made bells of pure gold...’ (Exod. 39.25)
In one verse, the text even repeats a noun in order to force the
words into a triple construct form:

DW "yy "Ta nara? DTA Myy
‘And thou shalt make staves for the altar, staves of acacia-
wood...’ (Exod. 27.6)

In this case, the second "72 is not necessary; it appears here
merely in order to present the material as a complement and not
as governed by mwyn.

In the instances presented thus far, the syntactic status of the
material is clear from the text itself. Sometimes, however, it is not
possible to determine the status on the basis of the text. For example:

+ TO ANT MBS wy

‘And thou shalt make an ark-cover of/with pure gold...’

(Exod. 25.17)

+ THO ADT PR wy)

‘And thou shalt make a plate of/with pure gold...’ (Exod.

28.36)

In some cases, the matter was determined by the vocalisers and
the cantillators in harmony with one another:

250 Breuer

aT nipay tw mribyy

‘And thou shalt make two rings of gold...’ (Exod. 28.26, 27)

DBO ATT HP yy

‘And thou shalt make unto it a crown of gold round

about...’ (Exod. 30.3)

TEN? NWHI ia] WH Wha Moy

‘Thou shalt also make a laver of brass and its brass base, ...’

(Exod. 30.18)

JT npav baw ayy

‘And they made two rings of gold...’ (Exod. 39.19, 20)

In other instances, the verses appear to unambiguously fa-
vour one interpretation, yet despite this both the vocalisers and

the cantillators adopted an alternative interpretation:

+08 TPA VY NUNIT NS Wy")

‘And they made the tunics of fine linen with woven work...’
(Exod. 28.39)

. Wn 132 N81 NwWHI Wan NY wy

‘And he made the laver of brass and its brass base,...’ (Exod.
38.8)

..Wy nbaya7 NE"NN) WY npivan hy

‘And the mitre of/with fine linen, and the goodly head-tires
with fine linen...’ (Exod. 39.28)

In these instances, the product is in the definite state while the
material is in the indefinite; nevertheless, both the vocalisers and
the cantillators perceived the forms as construct forms, contrary
to conventional grammar.

In some cases, verses that are virtually identical are treated

differently. The differences take the form of alternate readings of

Dissonance between Masoretic Accentuation and Cantillation 251

the biblical text itself or of the vocalisation or cantillation. Com-
pare the following verses:

70 ant A aANNY wy

‘And he made the candlestick with pure gold...’ (Exod.

37.17)

2 TO ANT Ma wy

‘And thou shalt make a candlestick of pure gold’ (Exod.

25.31)

DOW "ZY DIMI yy)

‘And thou shalt make bars with acacia-wood’ (Exod. 26.26)

DOW (YD TI Wy

‘And he made bars of acacia-wood...’ (Exod. 36.31)
In the above sets of verses, the alternatives in each pair are ex-
plicit in the form of the text itself. Elsewhere, the differences are
manifest only in the cantillation:

aT neawn snw wy

‘And they made two settings of gold...’ (Exod. 39.16)

aT ngawn myy

‘And thou shalt make settings with gold...’ (Exod. 28.13)

In other instances, the alternatives are only in the vocalisation:

DW "ED INPBITNS wy

‘And he made the table with acacia-wood...’ (Exod. 37.10)
DOW rey NW Ivy

‘And thou shalt make a table with acacia-wood...’ (Exod.
25.23)

DOW "8D DNTTNN ON?¥A Dyn

‘And Bezalel made the ark with acacia-wood...’ (Exod.
37.1)

252 Breuer

“DRY ZY TS WY
‘And they shall make an ark of acacia-wood...’ (Exod.
25.10)

The word jn>w ‘table’ appears consistently as an independent
word in Exod 37.10 and 25.23. By contrast, 7i7& ‘ark’ is presented
in the construct state in 25.10. Although it is impossible to deter-
mine the matter by cantillation, the vocalisers decided that the
form is a construct state. The net result is that this verse differs
both from the analogous verse about the ark and the similar
verses about the table.

In summary: the verb nwy is very common in verses de-
scribing the construction of the Tabernacle, and it appears along-
side the material from which the relevant item in the Tabernacle
is made: wood, copper, gold, or marble. We find two customs
regarding the relationship between nwy and the material: some-
times they appear as complements and in other cases not. The
distinctions between these two approaches are sometimes re-
flected in the biblical text itself, while elsewhere they are implied
by the vocalisation or cantillation. The custom of attaching the
material to nwy as a complement follows the approach of the can-
tillators in our case: JAX WInN-nwyA ‘an engraver’s work in stone’.
The alternative custom mirrors the approach of the vocalisers
here: jax-win nwyn ‘the work of a stone-engraver’. The only dif-
ference between this verse and others is that in this instance, the
Masora presents both approaches in a single verse, through the

vocalisation, on the one hand, and the cantillation, on the other.

Dissonance between Masoretic Accentuation and Cantillation 253

(2) :DAR? MAYAN TAI PANS OVINE TW TD
‘Behold, I have given him for a witness to the peoples, a
prince and commander to the peoples.’ (Isa. 55.4)

The cantillation here establishes the following division:
DAN) MINI 3

According to this division, nx ‘peoples’ is not part of a con-
struct, but complements the verb rnni ‘I have given him’. The
word must then be analysed as the preposition -) before the noun
orax ‘peoples’, reflecting the plural not of 0&>, but of AMX. Ac-
cordingly, neither 7°33 nor 71¥/) are construct forms.

However, the vocalisers pointed 71¥/) with a sere, reflecting

the construct state. According to this approach, the division is:
DAN MANO = Th

In this case, 07789 does not represent -9 + oAx, but is the plural
form of OX.

The Aramaic translations appear to analyse the verse in the
same way as the cantillators: 52 >y w5wi qn jw xnnyd 27 RA
xma5n ‘Behold, a leader for peoples I appointed him, a king and
ruler over all the kingdoms’ (TJ); ple .nrksou eons) wimw Xm
whan esinxma ‘Behold, a witness to peoples I made him, a ruler
and leader to peoples’ (Pesh.). The change here in the second
word (xm125n or chop, rather than xXnny/<sp.) implies that
they read -) + orax, as does the cantillation.

Qimhi also follows the cantillation, and accordingly ques-
tions the vocalisation: 37109 xXdw ."4¥1 MINA) ‘nIva with a sere,

contrary to custom’. Luzzatto is more ambivalent:

254 Breuer

ynni myn Pal DAR ww wil ...01xd ww > aR pwRoin
Daw paw pan 31 391305 TINT ,NDIND *NdA Mn Nom ,ony>
DAX WW oIwWn Ik OXd wIwnA

The first instance appears to me to be from the root ox)...

and the second from the root ony, i.e., I have given him a

prince and commander to the peoples; the word mii is

not in the construct state and should properly have a segol;

but it is also possible that both of them come from the root

ox) or that both come from the root onx.

Neither of these commentators mentions the fact that both these
possibilities are present before us in the masoretic form: one in
the vocalisation and the other in the cantillation.?

It is difficult to decide which reading represents the correct
interpretation of this verse. It only remains to add that according
to the division reflected by the cantillation, the two legs of the
parallelism contain two different words pronounced identically
(or”AX? and oAKX+)), and this may be regarded as poetic refine-
ment and elevation. This consideration is not decisive, but it adds
credence to the division of the cantillators.

(3) rma ban ope biPaa wT ATE Vp yA TaN! DAarny 'n TWD

DIP Tigw

‘For the LorD spoileth Babylon, and destroyeth out of her

the great voice; and their waves roar like many waters, the

noise of their voice is uttered.’ (Jer. 51.55)

Tixw ‘noise’ is vocalised here as a construct form, and the vocali-
sation presupposes the following division:

Dp piRw nl

3 See also M. Breuer (1982, 386); Hakham (1984, 590, n. 23).

Dissonance between Masoretic Accentuation and Cantillation 255

The subject is ‘the noise of their voice’, and as Rashi interprets
onpyy dip ynw: ‘the sound of their shouting was heard’.
However, the cantillation presents a different division:

Dip PRw {NI

The division in the cantillation guides us to the reading jixw*, in
the absolute state. Accordingly, the subject is o3ip ‘their voice’
alone, while ;ixnw* is not part of the subject, but an adverb com-
plementing mi ‘is uttered’. According to the cantillation, the
meaning of the phrase is ‘their voice is uttered noisily’, that is
‘their voice rose in volume’, or, as Qimhi explained: pixw 13 7M
pdip ‘And their voice was like a noise’. The Targum follows a sim-
ilar approach: pa>p xwiinsxa pam ‘and they raise with noise
their voice’. pnp is the object, while xwiinx2 translates nxw.
Again, therefore, ;ixw serves as an adverb, as implied by the di-
vision in the cantillation.*

The difference between the two interpretations of this verse
centres on the question as to whether ip and nxw have identical
meanings. If they do, their combination may be regarded as an
instance of hendiadys—np-pxv; the inverse version of this com-
bination—jixw->ip—appears elsewhere in the Bible (Isa. 13.4;
66.6). The vocalisers (and those who follow their approach) in-
terpret our verse in this manner. If the meanings of the two words
are not identical, however, then we do not have a single phrase
here. Since they are not identical, the sound (517) can be under-

stood to be made in a manner similar to a pw. In this reading,

* Wickes (1887, 68) rearranges the cantillation marks to suit the con-
struct form; see also Ginsburg (1926a, 207).

256 Breuer

pxw carries its own distinct semantic weight that is capable of
both describing ‘ip and adding to it something it does not carry

on its own.

2.2. Absolute in the Vocalisation, Construct in the

Cantillation

(4) oppr iors Tina api Snap Dw Taga PAN TS “IT ApS 1979
aywywn praqa oy) xban Toop
‘For thus saith the LorD: Behold, I will extend peace to her
like a river, and the glory of the nations like an overflowing
stream, and ye shall suck thereof: Ye shall be borne upon
the side, and shall be dandled upon the knees.’ (Isa. 66.12)

The prevailing interpretation among the commentators is ‘be-
hold, I will extend to her like a river—peace; and like an over-
flowing stream—the glory of the nations’. According to this ap-
proach, ‘like a river’ is a description parallel to ‘like an overflow-
ing stream’, while ‘peace’ is an object parallel to ‘the glory of the
nations’. Accordingly, this interpretation adopts the following di-
vision:
mw miaa—Avw>nia =o odw —an

This division is followed, for example, in the Peshitta: “ni “nm
ion Uo resale culs ein’ ‘Behold I cast upon her peace like a
river’. The Sages appear to have shared this understanding:
Avon A ax AD "nn jobw a> miod Ty cuxw anK — mndwn
pidw ania mx ‘nabwzn7 — refers to the nation to which I shall ex-
tend peace, as it is written: ‘thus saith the LorD: “Behold, I will

extend peace to her like a river”’ (Genesis Rabbah 66.2).

Dissonance between Masoretic Accentuation and Cantillation 257

However, Luzzatto already noted that the cantillation does
not lead to this interpretation. According to the above reading,
nibw should have carried an accent of the third degree (tevin),
which in this position would have been stronger than the tevir on
"vw. The geresh over nibw shows that the cantillation actually

adopts the following division:
oad yyw dni ow 3719

According to this division, there is only one object in this verse:
D134 7129. ow is not an object, but describes 773, and the two
images we have here are ‘as a peaceful river’ and ‘as an overflow-
ing stream’.

This is certainly the meaning intended by the cantillators;
it remains for us to ask only what the vocalisers intended. In
other words, which of the two above-mentioned meanings is im-
plied by the vocalised text before us: n1bw 7732?

Since ow is a noun, and not an adjective, whenever it de-
scribes the preceding noun it must form part of a construct. For
example: oiyw 127 ‘words of peace’ (Deut. 2.26); ni>w 72x91 ‘mes-
sengers of peace’ (Isa. 33.7); oiw nyu ‘and the counsel of peace’
(Zech. 6.13). And to quote examples when the construct state is
dependent on the vocalisation: ni)w 7132 ‘in a peaceable habita-
tion’ (Isa. 32.18); Dibw niawnn ‘thoughts of peace’ (Jer. 29.11);
Dow ritn ‘vision of peace’ (Ezek. 13.16); ni>w vawmi ‘and judgment
of peace’ (Zech. 8.16). Accordingly, if p1»w indeed describes 773,
the proper vocalisation here would be n1>w 1733; the vocalised
form o1>w 7732 is, therefore, not a possible reflection of this in-

terpretation.

258 Breuer

Accordingly, the vocalised text before us—nibw 77312—
clearly indicates that this is not a phrase consisting of a noun and
an adjective, but rather an adverb and an object. In other words,
and in keeping with the opinion of the commentators as we

quoted at the beginning of our discussion:
ou maa- vw ona = ow 4719

Thus, we see that the difference between the vocalisers’ ap-
proach and that of the cantillators centres on the interpretation
of ow, and hence also of 7719. The vocalisers read o1>w 7732 and
regard oibw as an object, i.e., ‘I will extend peace to her like a
river’; the cantillators read now 7732 and understand ‘peace’ as
describing ‘river’, i.e., ‘I will extend to her as a river of peace’.
(5) sane oryta casi 2 ayy OTN 1

‘A man’s gift maketh room for him, and bringeth him before

great men.’ (Prov. 18.16)

According to the cantillation, the division here is:

Qo DTN 119
The meaning is that a man’s gift will expand him or make room
for him. This is reflected in the Targum: xn RWI 737 ANANNIN
m5, and Ibn Ezra also interpreted these words in the same man-
ner.

However, jf is vocalised with a qamets, indicating the ab-
solute rather than construct state. Accordingly, the vocalisers did
not see a construct here, but rather interpreted jm” as focal. Ac-
cordingly, their division is:

bamIanvow wn

Dissonance between Masoretic Accentuation and Cantillation 259

And the meaning can be represented with a cleft sentence: ‘it is

a gift that makes room for a man’ (Eichel 1790, first introduction).

3.0. Definiteness

(6) 9 :9792 neva 199A ov IPT TAN nna
‘He opened the rock, and waters gushed out; they ran, a
river in the dry places.’ (Ps. 105.41)

The most probable division is that proposed by the vocalisation:
73 niva idn

The commentators understood the verse according to this divi-
sion. However, this creates a disagreement between the singular
subject 173 ‘river’ and the predicate’s plural verb 1299 ‘ran’. Vari-
ous solutions can be proposed in this respect. Some commenta-
tors suggested that the noun 0° ‘water’ (plural in Hebrew) also
governs the second part of the verse, so that the meaning would
be 791 7 niya 1990 ‘the waters of the river flowed in the dry
places’ (Ibn Ezra). Others argued that the singular ‘river’ here ac-
tually stands for a plural, so that the meaning is ‘the rivers flowed
in the dry places’ (Rashi). Still others opined that the meaning is
‘(the above-mentioned waters) flowed in the dry places like a
river’, e.g., NIM Tn XMInva 19°97 (Targum). In any case, this in-
terpretation clearly understands ni*y as the plural of nx ‘a wil-
derness’ (the only occurrence of a plural form of this word in the
Hebrew Bible).

The cantillation, however, divides the verse in a different

manner:

sai nyya adn

260 Breuer

What meaning is implied by this division? It would seem to be
founded on an interpretation proposed by the Sages, who suggest
that nivy is not the plural of 7°, but rather of x ‘ship’. The Tosefta
comments: 5xx at PRA MIXPAORA paw 77... 173 OOM Mwy xT
qm nya 157 'w ,nt ‘And it becomes mighty rivers... they sit in
ships and come to each other, as it is written: 173 nvya 1057’ (t.
Sukkah 3.12); and more explicitly in Bemidbar Rabba: wx
15m ‘aw jaraoa nodan Amn 5at> Sata Aman oxs 75% Alay annw
way xd oR ol Raw jniPaD NOX NPY pR1 WI nvya ‘a woman who
needed to go to her friend from a tribe to another tribe would go
by ship, as it is said 173 nvva 125m, and nvy means ships, as it is
written: 172" 85 77x »x1 ‘a mighty ship cannot cross it’ (Bemidbar
Rabba 19.26).° According to this interpretation, 171 nvy is a con-
struct chain, and the meaning of 171 nvya 1957 is ‘they went in
riverboats’.

However, according to this interpretation, nity cannot be
definite. In other words, the interpretation implied by the cantil-
lation requires a bet vocalised with shewa, rather than the patah
reflecting the definite article according to the received vocalisa-
tion (Hakham 1981, 277).

(7) THN ON iT NY wy raINyT
‘Did not He that made me in the womb make him? And did
not One fashion us in the womb?’ (Job 31.15)

The cantillation establishes the following division:
Tnx onja int

5 See also Tanhuma Huqat 21. And see Ben-Yehuda (1908-1958, 5646a,
n. 1).

Dissonance between Masoretic Accentuation and Cantillation 261

According to this division, the phrase here is 7nx on ‘one
womb’, whereby ‘one’ describes ‘womb’. Since Tnx is not definite,
on must also be construed as indefinite, and accordingly the
cantillation requires the reading onn3, with bet vocalised with
shewa. Several commentators interpreted the verse in this man-
ner, including the translator of the Peshitta: ash Rssim ass.

However, the vocalisation gives us the definite form 0m72,
thereby establishing that we do not have a single phrase here.

According to the vocalisation, the division is as follows:
TNX onqa 139

This interpretation makes 1nx the subject of the sentence, rather
than a complement of om, as interpreted by Ibn Ezra, among
others: 7n& 5x oma 1119" ‘One God fashioned us in the womb’
(Norzi 1742-1744, VI:50a; Hakham 1970, 238, n. 91; Qafih
1973, 263).
(8) saw? Dawa opwyI Oa OATS 9200 1A}
‘Folly is set on great heights, and the rich sit in a low place
(Eccl. 10.6)

The division according to the cantillation is:
Taw Dawa oywy oman ona DOF NI

Here 0°35 on” is a phrase (‘great heights’) whereby 0°27 de-
scribes 0°10N1. This interpretation was followed by Rashi, for
example: 71213 ANNA ywrni niwwn yn Iw ‘For foolishness and evil
was placed at a great height’. Once again, the syntactic difficulty
is glaring: the noun is definite while its adjective is not. However,

the definite character of o°217N1 is indicated not in the written

262 Breuer

text itself, but in the vocalisation. Therefore, it seems that the
cantillators read 0°37 071703, as an indefinite construction.

By establishing the definite vocalisation n°nin”a, the vocal-
isers disconnected the two words, resulting in the following divi-

sion:
yaw Saws o wy 04 nina S907 jn

According to this interpretation, 0°29 is not an adjective
complementing o’amna, but rather joins the second half of the
verse, alongside o-wyi. Ibn Ezra accordingly comments: w1"51
ov >173 199 Dra ‘And the meaning of om is like (that of) D173
“great ones” (Yalon 1971, 331).

Lauha (1978, 183) adopted a similar line: “Nach
masoretischer Akzentuierung Gehort 0°25 zu V. 6a, aber wegen
seiner Artikellosigkeit kann es nicht als Attrubut zu mann.
gehoren und muss mit V 6b kominiert werden” (‘According to the
cantillation, 0°25 belongs to the first part of the verse. However,
since 0°17 is indefinite, it cannot serve as an adjective of o*nNnA
and must join the second part of the verse’).

These comments are very pertinent but require a slight cor-
rection. The indefinite character of 0°." indeed disconnects it
from o7210n1—if that word is definite. But its definite character
is conveyed solely by the vocalisation. It is thus evident that by
joining the two words together, the cantillators read not naira,
but rather n’ninaa, and according to this reading there is no dif-
ficulty in connecting the two words. The dissonance here, then,

is not one between the cantillation and the rules of grammar, as

Dissonance between Masoretic Accentuation and Cantillation 263

Lauha’s comments may imply, but merely one between the can-
tillation and the vocalisation (see also Zer-Kavod 1973, 63, n.
15).

4.0. Pausal versus Non-pausal

4.1. Introduction

This section will discuss exceptions to the usual rules concerning
pausal and non-pausal forms in the Bible. As a general rule,
pausal forms appear with strong disjunctives, while non-pausal
forms accompany weaker disjunctives and conjunctives. Clear ex-
ceptions to these rules can sometimes be found, however, and
these can be explained according to the underlying message of
this article: the vocalisation may be exceptional because it is
fixed according to a different division.

In order to do so, we should mention some basic fundamen-
tals regarding pausal and non-pausal forms:

(1) A vowel that stems from an original i vowel does not
change in pause, as *nipt ‘I am old’ (Gen. 27.2); ox ‘softly’ (1 Kgs
21.27); naa ‘as a daughter’ (2 Sam. 12.3) (Breuer 1980, 244-46;
Blau 1981). Sometimes a vowel does not change when its origin
is not certain, such as 11758 jin ‘all kinds of stuff (Ps. 144.13);
4x90 ‘the steward’ (Dan. 1.11); 527 ‘Michal’ (1 Sam. 14.49).
Such forms cannot be considered non-pausal forms, as they do
not change in pause.

(2) Pausal forms may appear with accents other than the
chief disjunctives, e.g., Tvina NoaT NY OXI ww ‘let the chil-
dren of Israel keep the Passover in its appointed season’ (Num.

9.2). Consider, for example:

264 Breuer

228 PAM TW TWIN) WRB TI? ART pT

‘Benjamin is a wolf that raveneth; in the morning he de-
voureth the prey, and at even he divideth the spoil.’ (Gen.
49.27)

We may be tempted, according to the forms, to propose a

different division:
Sw pom any ty axe apa. = wy aNT pva

While this division may indeed be more plausible, we cannot
reach this conclusion due to the considerations discussed above.
On the one hand, a pausal form of the type found here—yAw—
may appear with a strong disjunctive, such as the first zagef in a
verse. On the other hand, it is doubtful whether 7p is truly a non-
pausal form, since there is no instance of this word vocalised with
a games in a pausal context (pace Schlesinger 1962, 88-89).

(3) There are various types of pausal forms, and each type
behaves differently. Some types readily tend to adopt the pausal
form, while others do so only with the chief disjunctives. Let us
take the following example (the division here follows the cantil-
lation):

TPH IVY —- TVA yAIPH 73a yw

‘When thou sittest in thy house, and when thou walkest by

the way, and when thou liest down, and when thou risest

up’ (Deut. 6.7)

It is clear that the only reasonable division of this verse is
according to the cantillation. However, we should not compare
the pair of forms 7m2/7mA1 with the pair 777/777, since the ten-
dency to the pausal form is extremely weak in the latter type.

The situation is further illustrated by the following example:

Dissonance between Masoretic Accentuation and Cantillation 265

TNA AWIN?) DDR Twa ays NX? TINA RITMIay

‘Let us pass, I pray thee, through thy land; we will not
pass through field or through vineyard, neither will we
drink of the water of the wells.’ (Num. 20.17)

Based on the forms in this verse, we should ostensibly depart
from the cantillation and divide this verse as follows:

SRAM ANWIRNd OW ATWATAyIRD FINA NITY]

Here, too, although this is a more plausible division than that
established by the cantillation, we cannot claim that it is dictated
by the vocalisation. The pair 7¥78/7¥78 cannot be compared to
the pair 072/072, since the former is far more prone to adopt the
pausal form than the latter (Ben-David 1995, 8-9).

Accordingly, when comparing pausal and non-pausal forms
in the same verse, we must draw conclusions only if the com-

pared forms belong to the same type.

4.2. Pausal Form in Context

(9) nin Sapa ken WRa NZowWeT OU PENT Aa? DT? WN}
‘And I said unto them: “Whosoever hath any gold, let them
break it off; so they gave it me; and I cast it into the fire,
and there came out this calf.’ (Exod. 32.24)

In this story, Aaron tells Moses about the sequence of events that
ultimately led to the sin of the Golden Calf. His speech includes
an ambiguous verbal form 3p79n7. This word can be understood
in two ways, both grammatically and contextually: either as an
imperative form or as a third person plural past tense. If it is an
imperative, then it is part of Aaron’s words to the people that are

being quoted here, as he tells them ‘whosoever hath any gold—

266 Breuer

break it off. Thus Aaron is repeating his comments as reported
earlier in the story itself: >3n¢2 Wx AnH ATI PID FADS DAN WAN
25x 4x74) DIMI21 D"Ia O53 ‘And Aaron said unto them: “Break
off the golden rings, which are in the ears of your wives, of your
sons, and of your daughters, and bring them unto me”’ (Exod.
32.2). If this is a third person past tense form, then 271 "79 ‘who-
soever hath any gold’ are the only words here that Aaron actually
spoke at the time of the event, while the following section—
un’ ip yani—describes the people’s resulting action: ‘they
broke off and gave it to me’—paralleling the content of the story
itself: NON WP) DP INA Wy AD AInS DbT-YD Ipqann ‘And
all the people broke off the golden rings which were in their ears,
and brought them unto Aaron’ (Exod. 32.3).

According to the cantillation, the division of the verse is as

follows:
Sam ran aAtond on 7x

The meaning according to this division is clear: Aaron’s speech is
short, and confined solely to the words 171 °19, while 1p75n7 is a
past tense form describing the people’s actions, and combined in
the division of the verse with the adjacent *7-11n". Rashi followed
this interpretation: nn om 3725 ,'ant n>! stn& aT — ond TRI
> yam iprpni ‘And I told them — one thing only: “Who has gold?”
And they hurried and broke [it] off and gave [it] to me’.

This interpretation reflects a desire to limit Aaron’s speech
during this affair to the minimum, thereby also mitigating his sin:
he spoke only two words, and did not himself tell the people to
break off their gold. The main sin is that of the people, who broke
off their gold without having been told to do so. The desire to

Dissonance between Masoretic Accentuation and Cantillation 267

limit Aaron’s speech is not confined to the exegesis, however, and
can be seen in Aaron’s words as reported. Even if we extend his
speech to include the word 1p19nj, this is still markedly concise
by comparison to his words as quoted in the story itself: ‘12 3779
Pox WPT) DIMI 072 IW) Na WW Aq ‘Break off the golden
rings, which are in the ears of your wives, of your sons, and of
your daughters, and bring them unto me’ (Exod. 32.2). Accord-
ingly, even setting aside exegesis, it seems that Aaron was eager
to shorten his speech, for understandable reasons. The commen-
tators followed this trend, truncating Aaron’s quoted speech still
further.

However, the vocalisers pointed 1)75n7 as a pausal form. A
pausal form in this context, with a tippeha after a zagef (which is
the second king in a verse that also includes an ’atnah) is an un-
usual occurrence. Accordingly, they appear to have divided the

verse differently:
> um APAanA anreny ony 7K

According to this division, 1)78n7 is included in the quote of Aa-
ron’s speech: ‘whosoever hath gold—break it off!’ The people’s
actions are now confined to "yn" ‘and they gave it me’. This
interpretation also has support among the commentators. Avar-
banel, for example, explains: !prpnn—ant ond :om>x om yh)
Oya opynni ining Aly ann > ww AnD ann tnx 52 sand ayn
‘,..And so I told them: whoever has gold—take it off! That is to
say: each person will give gold according to how much they have,
those with much will give more and those with less, less’.

As noted, the thrust of Aaron’s comments leaves room for

both these interpretations. The same is true of the comparison to

268 Breuer

the original story in the text, which includes both an imperative
form 3p78 and the past form 3p7|n"; the form here may be con-
sidered to mirror either of these. If we turn to the plain meaning
of the verse, however, there seems to be no alternative but to
read 1p19n7 as an imperative form, for a simple reason. A past
tense form would be expected to appear in an inverted tense
form, as is usual in a narrative chunk and in keeping with the
surrounding past forms: 8¥"1 ...1729Wx) ...711n"1 ...099 Toki ‘and I
said to them... and they gave... and I cast it... and (this calf)
came out...’ (Exod. 32.24). The exceptional form here shows that
this is an imperative concluding Aaron’s words.

For our purposes, what is important is not the original in-
terpretation of the form, but the interpretation adopted by the
vocalisers and the cantillators, and in this respect they were di-
vided: the vocalisers interpreted 1)75nn as an imperative, while

the cantillators saw it as a third person past form.

4.3. Pausal Forms in Non-Pausal Contexts and Vice-

Versa

(10) 9p an ray had) WaT ON DUN? TIWwa NZI ON TR NAT
77 M203] VINNY NAT] AWW nw
‘And he shall bring a ram without blemish out of the flock,
according to thy valuation, for a guilt-offering, unto the
priest; and the priest shall make atonement for him con-
cerning the error which he committed, though he knew

it not, and he shall be forgiven.’ (Lev. 5.18)

Dissonance between Masoretic Accentuation and Cantillation 269

This verse describes the sacrifice known by the Sages as "bn oWx
‘contingent guilt offering’. The cantillation establishes the follow-
ing division:

1% nyon YTOND NIM WW InIw Op

The first part of this clause includes two grammatically similar
verb forms: 3xw and y7°. As noted above, such a situation is opti-
mal for purposes of comparison, since in such a case the forms
are equally prone to pausal formation and they occur within a
single verse. Yet as we see, the pausal form appears on the weaker
cantillation mark, while the principal mark carries a non-pausal
form. Accordingly, the vocalisation implies a different division:

1 ndon yt Xo xm 338 WR InwWw Oy

Thus the cantillation and the vocalisation disagree as to whether
the intermediate section y1~N> xim is to be attached to the first
part of this clause or to the second (Ben-David 1995, 154, n. 146).
What is the background to this uncertainty?

We should firstly note that the words yT-x> x17) stand out
here, since they do not appear in the analogous verses concerning

other sacrifices, such as:
7 don ingen 137 rev 731...
‘,..and the priest shall make atonement for him as concern-
ing his sin, and he shall be forgiven.’ (Lev. 4.26)
27 N20] NENTS INV THT MY II...

‘,,.and the priest shall make atonement for him as touching
his sin that he hath sinned, and he shall be forgiven.’ (Lev.
4.35)

270 Breuer

Accordingly, the inclusion of these words here is inherently prob-
lematic. On the one hand, it is possible to assume quite simply
that yt’ x5 xim is merely a reiteration in different words of inz3w
3x WR. This is surely the understanding implied by the cantilla-
tion in combining these two phrases.

However, the Sages saw in these words a specific reference
to the special law concerning this sacrifice, the contingent guilt
offering. They explain that contingent guilt occurs only when a
person is not certain that he has committed a sin. This offering
cannot atone for a sin unless the sinner is unaware of his sin and
as far as he does not have certainty. Even after making this offer-
ing, if he learns that he indeed sinned, he is obliged to make a
new sacrifice—the sin offering (nxon)—like a regular sinner: 19m
NTI pty ty > ndim paon Sy na2n gIAW an—"7dn OWRK KIPIA
INNON A 7p 733wa Xvnw ‘And this is what is known as the contin-
gent guilt [offering]—since it atones for doubt and remains pend-
ing until he knows with certainty that he sinned in error and he
makes his sin offering” (Maimonides, Mishneh Torah, Shegagot
5.1). Sifra (Finkelstein, 1983, 209) explains: ,1> noon yr xd xim
manyiw 7) by ax nanny adayd nan ar and xn jb 72Dn70 sx yp oR NA
Pom Aran sn KXvAI TD AnKi “And he knew it not and is for-
given”’—implies that if he did know it, he is not atoned. What
does this resemble? The red heifer, that if the murderer was
found after the heifer’s neck had been broken, he will be exe-
cuted’. In these comments, the expression y1-X> xim ‘and he
knew it not’ clearly refers to the post factum situation, and if he
later learns that he indeed sinned, his atonement is nullified, as
Rashi explains here: 7y mt owxa 1 7D9n2 XD aT IN&> pT DX NA

Dissonance between Masoretic Accentuation and Cantillation 271

nxon x2w ‘Yet if he knew later, he is not atoned for this guilt
until he has brought a sin offering’. According to the Sages, then,
forgiveness depends and is conditioned on the sinner not having
been aware of his sin; as soon as he becomes aware of it, the
atonement is nullified. This would seem to explain why the vo-
calisers formed the unit 1) non yT-x> xin, to emphasise that he

is only forgiven for as long as he is unaware.

(11) 'A7Ax> 'A owa unix Piw avaws awe oidwd 7) 7179 iNT TaN
‘DZ Ty WWM Pay pa Thy ra AM
‘And Jonathan said to David: “Go in peace, forasmuch as
we have sworn both of us in the name of the Lord, saying:
‘The Lord shall be between me and thee, and between my

seed and thy seed, forever.”” (1 Sam. 20.42)

The cantillation establishes the following division:

Day-Ty qr pavwar pa Thaops ja

However, the vocalisation of the forms here challenges this divi-
sion, since 73°21 is vocalised as a pausal form, while 7D71 appears
in a non-pausal form. Here, too, the two words belong to the
same category (shewa in the non-pausal form versus segol in the
pausal form in a second person masculine singular pronominal

suffix), and again they appear in a single verse. According to the
vocalisers, then, the division is:

Do TY A paryar pa yar wan

The vocalisers and the cantillators disagreed as to whether
ody ty in this verse is attached solely to Jynt 7721 °yot pa (as the
vocalisation dictates), or also to Jr21 71 (as the cantillation im-
plies) (Ben-David 1995, 72, n. 110).

272 Breuer

The question whether o>1y ty properly describes only the
following generations or also the current one arises in numerous
verses. In several places in the Bible, we find the emphatic parti-
cle oy used to strengthen yt ‘seed’. The following is one exam-
ple of such a verse: :o3iy ning ans qr? NNT] PINATNN Anh...
‘...and I will give this land to thy seed after thee for an everlasting
possession.’ (Gen. 48.4)

Conversely, no1y sometimes undoubtedly also refers to the
current generation, as in the verse shortly before the one we are
discussing here: :o%iy-ty qraiora ' mn... ‘...behold, the LorD is
between me and thee forever.’ (1 Sam. 20.23)

In several verses, however, the question involves the divi-
sion of the verse, and the commentator or cantillator must decide
what is complemented by the word ony. This is the case in our
verse, where the cantillation determined that od1y refers to this
generation, as well as future ones. The same is true in the follow-

ing verses:
DTI? OTT PINS Wt Pa TI Ps...

‘...between Me and thee and thy seed after thee throughout
their generations for an everlasting covenant...’ (Gen. 17.7)

oop) =o orpaadr ond avr war...

‘,..that it might be well with them, and with their children
forever.’ (Deut. 5.25)

Dy PA TTT? PwaA? TorAvY...

‘...and showeth mercy to His anointed, to David and to his
seed, for evermore.’ (2 Sam. 22.51)

DDiy-TyT AAyA 'A TAX AVI pata yw IT 81 PAN...

Dissonance between Masoretic Accentuation and Cantillation 273

‘...(they will not depart) out of thy mouth, nor out of the
mouth of thy seed, nor out of the mouth of thy seed’s seed,
saith the Lord, from henceforth and forever.’ (Isa. 59.21)

However, we also find a different pattern of division,
whereby the cantillation establishes that oi refers solely to the
seed:

DPW FN) TYAN 77 ART ARR WY PISA YT "2

‘for all the land which thou seest, to thee will I give it, and

to thy seed forever.’ (Gen. 13.15)

DPI WH pi NIN? 72 NT

‘and they shall be upon thee for a sign and for a wonder,
and upon thy seed forever.’ (Deut. 28.46)

OP WT UNS gh WNyR BDTV)

‘so shall their blood return upon the head of Joab, and upon
the head of his seed forever...’ (1 Kgs 2.33)

D2? WU TRPITA yay nv

‘the leprosy therefore of Naaman shall cleave unto thee,
and unto thy seed forever...’ (2 Kgs 5.27)

It emerges that the cantillation itself follows two ap-
proaches, sometimes establishing that 0>1y relates solely to the
coming generations and sometimes includes this generation. Ac-
cordingly, it is hardly surprising that in our verse, too, the Maso-
retes disagreed regarding the division, though in this instance the
disagreement was between the cantillators, on the one side, and

the vocalisers, on the other.

274 Breuer

5.0. One Word or Two?

5.1. m5axn, mnandw

(12) pri MER PIS OX RW Sn TaTAD TIT Wy] Ops AT
SPON Tiv RiarNid wT ay TAK
‘O generation, see ye the word of the LorD: have I been a
wilderness unto Israel? or a land of thick darkness?
Wherefore say My people: “We roam at large; we will come
no more unto Thee”?’ (Jer. 2.31)
(13) Dixgwa nwp naax Diaa nwa qwinr>y onina zad-ov onina usw
ZMNANW Wy BW TSW] ANIP
‘Set me as a seal upon thy heart, as a seal upon thine arm;
for love is strong as death, jealousy is cruel as the grave;
the flashes thereof are flashes of fire, a very flame of the
Lord.’ (Song 8.6)

The two words we shall discuss here, 755xn and mnanbdw, are
similar in two respects: (A) They both end with the same -yah
suffix; (B) The first part of the word, before the suffix, is also
found in the Bible as an independent word: b»xn (Josh 24:7),
nanbw (Ezek 21:3; Job 15:30). Needless to say, the suffix 7 is also
well-known as an independent word. Accordingly, the obvious
question regarding each of these two words is whether they con-
stitute a single word, lengthened by the suffix, or two separate
words joined as a construct form.

m5axn: this form is usually considered a single word. Opin-
ion is divided as to its precise meaning—‘darkness’ or ‘wilder-
ness’—and regarding its vocalisation—n xn or TANN (see, for

example, Ginsburg 1880-1885, 602). However, there appears to

Dissonance between Masoretic Accentuation and Cantillation 275

be a consensus that this is indeed a single word. The cantillation
mark under the pe is not a tippeha and cannot be considered such,
since a disjunctive can never occur in the word of another dis-
junctive. This mark is considered a me’ayla, and so it is recorded
in the Masora Magna at Num. 28.26 (Breuer 1982, 106, n. 39).
The me’ayla has the same appearance as the tippeha, but marks
the place of secondary stress in the word and only ever appears
in words marked with silluq or ’atnah.

mnanbw: some of the commentators and translators regard
this as a single word, such as the Peshitta: ~susm\e. Others read
two words here; the Targum, for example, has: xnw x7 paw) p07
‘aM KIT ora ‘they are like coals of the fire of hell which God
created’. Ibn Ezra commented: 75n x°7 OX TIDAA wx pa npdna
5x 02 109 own ‘pnd NW XyNW ANpni .ornw ix nnx ‘There is a
dispute between the Masoretes as to whether this is one word or
two. Most probably it is two, and the construct form with the
Divine name is as in 5x 1772’. See below for other testimonies
concerning this word (Ginsburg 1926b, 574).

These two words resemble others that raise a similar prob-
lem. We will briefly review what is known about these other
words from the writings of the Sages. The sources for the discus-
sions about these words are of three types: (a) lists prepared by
the various Masoretes; (b) the Masora of the Targum; (c) the Tal-
mudic discussion (b. Pesahim 117a).

The following are the disputed words, with the identifica-
tion of the disagreeing parties:

mo2 ‘throne of the Lord’ (Exod. 17.16 [L]): Western/East-
em (Ginsburg 1880-1885, 1:592, 709,

276 Breuer

I11:191; Yeivin 1968, 80-82); School of Sura/School of
Neharde‘a (Yeivin 1968, 80-82; 1980, 121-22); Amoraim
in the Talmud

mort ‘Jedidiah’ (2 Sam. 12.25 [L]): Western/Eastern
(Ginsburg 1880-1885, I:593; Yeivin 1968, 80-82); Amo-
raim in the Talmud

mm%5n ‘hallelujah’ (Ps. 104.35 [L]): School of Sura/School
of Neharde‘a (Norzi 1742-1744, IV:22b—23a [Ps. 104.35];
Ginsburg 1880-1885, I:709-10; Yeivin 1968, 80-82)

manna ‘with great enlargement’ (Ps. 118.5 [L]): School of
Sura/School of Neharde‘a (Ginsburg 1880-1885, III:191;
Yeivin 1968, 80-82)

mnanbw ‘a very flame of the Lord’ (Song 8.6): Ben-

Asher/Ben-Naftali (Ginsburg 1880-1885, I11:191; Lip-

schtitz 1965, 53)

Thus we have located five words where there is disagree-
ment as to whether we should read one or two words (Yeivin
1968, 80-82). In addition to the sources quoted above, we must
now consider how these words were presented in the biblical text
(Ginsburg 1966, Introduction, 375ff; Yeivin 1968, 80-82):

moa / mt o3
Za he
Ambon

ay ayaa
mnayw

If we look at these words as presented here, we can discern
three different sources of opinion for the manner in which they
are to be perceived: spelling, vocalisation, and cantillation. In re-

gard to the spelling, is the word written as one connected word

Dissonance between Masoretic Accentuation and Cantillation 277

or as two separate ones? In regard to the vocalisation, does the
he have a mappiq (implying that the latter part is perceived as the
Divine name and hence the entire form is two words) or does it
lack a mappigq (in which case the latter part is not perceived as
the Divine name and the form is considered a single word)? In
regard to the cantillation, does the word carry one cantillation
mark or two? The following are the perceptions of these words

according to these three sources:

Tradition One word Two words
Spelling (2?) MoD (?) mM 02
mop m 57
mnandw m ann.
Vocalisation = mrnan>w mm 0D
mp
mada
ay anya
Cantillation m0
vey
aoa
manna
mnanw

Thus, we can see that the doubts detailed above were not
settled in a uniform manner in the final form of the biblical text.
The varying approaches led to the emergence of a mixed system,
including not only differences between the words, but also be-
tween the three foundations—spelling, vocalisation, and cantil-
lation. It may be worth adding that the cantillation is the only
one of these three foundations that adopts a consistent approach

to all five forms, interpreting them as two words in each case.

278 Breuer

From the standpoint of our subject here—dissonance be-
tween the vocalisation and the cantillation—only one of the
words is relevant: 77nandw. As discussed here, the vocalisers fol-
lowed the spelling in reading a single word, and accordingly
there is no mappigq in the he. The cantillators clearly took the po-
sition that this form constitutes two words, and accordingly it
bears two cantillation marks.

We may now turn to the second of the words on our list:
mbaxn (Jer. 2.31). This form is not mentioned in the various dis-
agreements between the Masoretes, and ostensibly all agree that
it constitutes a single word. This is supported by the above ex-
planation, that this form bears the rare cantillation mark me’ayla,
which only appears in a word accented with a chief disjunctive.
However, we must raise some questions in this regard. Firstly,
even if we have not found disagreement among the Masoretes
regarding this form, is it really impossible to understand it as
constituting two words? After all, and as explained, this form also
visibly comprises two words that exist independently in the Bible.
And more than one commentator has indeed sensed that while
this is one word (according to their perception and as determined
by the spelling), it actually comprises two words:

> Sy ax yong aon xen aon Soa—'a Tawa WANN pix Dax
Tar osx ain tot oy ya barn ands ,mdn ‘a aripw

But 75 ax” pix in Jeremiah 2—in all the books it is a single
word, despite the fact that its meaning is as two words, vis.
m 5axn, similar to 5x "11n and so forth. (Sefer Shewa‘ She-
mot, in Ginsburg 1880-1885, III:191)

Jodi .enandw ya mda onw ary nny ada xen onaon bom
ana mnanbw qnow ino ,baxn ana qram own ox bora

Dissonance between Masoretic Accentuation and Cantillation 279

7979 192 ,own Dx imix Taio etand Ayw Ia 52 yn .nandwa
py I23 ,oNdx Minar jobs ntInd cam ,ondNd Adi Ty ,ON
"apd

And in all the books this is a single word, though its mean-

ing is as two words, and the same is true of Wnandw. And

the construct of box with the Divine name emphasises the

darkness, just as the construct mnanbw emphasises the

flame. And thus when one wishes to amplify anything it is

placed in construct with the Divine name, such as 5x "1172

‘as mighty mountains’, on>xd 75173 Wy ‘an exceeding great

city’, oondx nt1nd nn ‘so it grew into a great terror’, *1n53

on>x ‘mighty wrestlings’, 'n 7155 Tx 123 ‘a mighty hunter’.

(Qimhi, in Biesenthal and Lebrecht, 1847, 25, discussing

the root 5"5x)

Secondly, we must note the unusual cantillation here: this
is the only place in the Bible where the me’ayla is preceded by a
tevir and a merekha, accents that are usually subordinate to a
tippeha (Breuer 1982, 106). This abnormality may ostensibly be
resolved by noting that a me’ayla also behaves like a tippeha in
an additional respect: in several instances it is preceded by a
zagef as the ultimate king before the chief disjunctive, and there
is no tippeha at all. An example of this is n2a"pNa oONDAN OPA
panvawa ‘2 AWIn ANN ‘also in the day of the first-fruits, when
ye bring a new meal-offering unto the LorD in your feast of
weeks’ (Num. 28.26). This phenomenon is found solely before a
me’ayla. As has already been observed (Breuer 1982, 106), the
melody was surely similar to that of the tippeha (as the graphic
form also implies), and so it was possible to place a zagef before
it by way of a final king, despite the fact that the final king is
always a tippeha.

280 Breuer

However, this deviation is not particularly serious. A zagef
serving as a final king is subordinate to the chief disjunctive, as
befits a king. The issue is only that it has not become a tippeha as
is expected with the final king. The quandary may be resolved by
means of the musical cantillation. Since the melody of the me’ayla
already provides a fine preparation for the ’atnah, there is no
longer any need to replace the zagef with a tippeha. In other
words, the zagef and the tippeha enjoy equal status, and the dis-
tinction between them is a purely melodic one. Accordingly, ex-
changes between these two marks may be explained on melodic
grounds.

In our current case, however, the quandary is vastly more
serious, since we have here a serious deviation in the basic rules
of the cantillation. The dominion of the ’atnah contains only a
single disjunctive, namely a tevir, so the dominion of the first-
degree disjunctive is divided by a third-degree disjunctive. Yet
the basic rule of cantillation, from which all the other rules and
details derive, is that the dominion of the disjunctive is divided
by means of a disjunctive which is one rank lower. To the best of
my knowledge, there is no other example of such a deviation in
the Bible—regarding not only ’atnah, but all the cantillation
marks.

Above all, however, we must pose the following question:
How can we be sure that this cantillation mark is indeed a me’ayla
and not a tippeha? Let us recall that the sole evidence that this is
indeed a me’ayla and not a tippeha is the fact that it appears on
the word of the chief disjunctive, and we can find no other tippeha

in this slot. However, the assumption that the cantillation mark

Dissonance between Masoretic Accentuation and Cantillation 281

indeed appears here on the chief disjunctive word depends on the
assumption that this is indeed a single word. If we assume that
these are two separate words—7 5xn—then this mark does not
appear on the last word, and, accordingly, must be a tippeha and
not a me’ayla, since a me’ayla appears solely on the last word.
Thus, the interpretation of this mark as a me’ayla and the inter-
pretation of this form as a single word are mutually dependent:
those who regard the form as a single word must argue that the
cantillation mark is a me’ayla and vice versa; and those who per-
ceive the form as two words must assert that the mark is a tippeha
and vice versa.

If we recall our comment above that the cantillators con-
sistently regarded all the words discussed here as two separate
words, it is not difficult to hypothesise that here, too, the cantil-
lators were faithful to their method. They perceived this form as
two words and marked it with a tippeha and an ’atnah, which are
preceded by marks customarily subservient to a tippeha.

As the masoretic form took shape, a situation emerged
whereby this word, like others, reflected a blending of contradict-
tory approaches: in this instance, spelling and vocalisation as one
word, on the one hand, and cantillation as two words, on the
other. Later Masoretes who saw this as a single word (under the
influence of the spelling and vocalisation) could now only define
the cantillation mark here as a me’ayla, since a disjunctive never
appears in the same word with another disjunctive.

If we accept this assumption, all that remains is for us to
discuss the meaning of this verse according to the cantillation.

First, we must discuss all the words perceived (by anyone) as two

282 Breuer

words. Even if the word is perceived as two words, the second
component—n’—may be in construct with the preceding word,
so that the difference in terms of the reading is not great. This is
the case, for example, with the words 77 177", 7 0D, and 7 nandw.
However, once the form is perceived as two words, this creates
the possibility of a different verse division, and of regarding 7”
not as a suffix, but as filling some other grammatical role. This is
what happened with the word 71m7n:

TT, VS AY AON]? WATT
‘Out of my straits I called upon the Lord; He answered me

with great enlargement.’ (Ps. 118.5)

Those who perceive 71n7/ as a single word will, naturally, re-
gard the entire phrase as a descriptive complement. Those who
regard the form as two words may also interpret it as a construct
filling the same role, in which case the division remains the same.
However, since they regarded the form as two words, the cantil-
lators went one step further, understanding yah as the subject of

the sentence, so that the division is:

mv anal 1p
rather than

mv ania Wy

What happened in our verse? Here, too, we may regard the
form as comprising two words, but in a construct form, as Qimhi
notes in his commentary: nan>w 71,0" nod ims Jad Dax aad
ods naan vam jo nboRd adits wy ,ox sna ,n") ‘And to magnify
the [concept of] darkness, he attached it to the word 7, as in
m nandw, dx sano, ondxd noi Wy, O'TDON MTINd cnn’. However,

Dissonance between Masoretic Accentuation and Cantillation 283

the cantillators placed the tippeha (as we see it) on Dxn. Accord-
ingly, their division is:
Sew) nea atan
m™ DaNn pax oN
The meaning is thus: ‘Have I been like a wilderness to Israel? Or
is the Lord a dark land?’ In other words, 7 here is not in construct

with 5oxn, but is the subject of the sentence, paralleling *n”7n in
the first part of the verse.

5.2. Taran

(14) a0 aN? TMAIAN P72 72 THAD N|2N DTZ WR) ONT TAY IN|
TMAVTINY NWI NDI? NPR kang Da

‘Then Daniel answered and said before the king: “Let thy

gifts be to thyself, and give thy rewards to another; never-
theless, I will read the writing unto the king, and make

known to him the interpretation.” (Dan. 5.17)

Jmaran is vocalised as the plural form of 73123, which appears
once in Biblical Aramaic in a similar context: x*av 49") Naan nA
‘gifts and rewards and great honour’ (Dan. 2.6). However, our
word appears with two cantillation marks, a pattern that is not
found in similar circumstances. Accordingly, this cantillation
would seem to be appropriate for those translations that see two
words here, such as the Peshitta: Jn’. 7p’x). However, this read-
ing requires a division into two words and a change in the vocal-
isation: JN73 11H.

Thus, the cantillation and the vocalisation disagreed here.
The vocalisation understands a single word—the plural form of

narax; the cantillators see here two words—yma"133.

284 Breuer

6.0. Conclusion

The Masoretic Text of the Hebrew Bible was consolidated after
many years of inspection and examination, and the Masoretes la-
boured tirelessly to clarify and shape its form. They were re-
quired to resolve numerous disagreements, some of which re-
mained unresolved even toward the end of the masoretic era.

These disagreements emerge before us in the masoretic lists
or as textual variants. Sometimes, however, these disagreements
can be discerned within the final masoretic text. The qere-ketiv
alternates are the most prominent example of this phenomenon,
generally representing two ancient versions. The written and re-
cited forms were not finalised simultaneously, and we accord-
ingly find two different versions reflected in a single biblical
form.

In this article, I have attempted to show that a similar phe-
nomenon exists between the vocalisation and the cantillation:
two different words or interpretations found their way into the
biblical text before us and survived in their original forms. We
find them before us now in a single version, one in the vocalisa-
tion and the other in the cantillation.

The precise explanation for this phenomenon is not entirely
clear. We may assume that there were from very early on two
different traditions, each with its own vocalisation and cantilla-
tion. The different methods and numerous disagreements led to
the blending of these distinct traditions as reflected in the final
Masoretic Text. However, it may be that the cantillators were ac-
quainted with the vocalised form as we have it, yet nevertheless
cantillated the text differently.

Dissonance between Masoretic Accentuation and Cantillation 285

Even if the precise reasons behind this phenomenon have
not yet been fully clarified, its recognition is nevertheless im-
portant. First, in regard to the interpretation of the verses, com-
mentators who strive to utilise all the tools in their possession
must not only consider the vocalisation and the cantillation to-
gether, but must also examine each separately and determine
whether they are compatible. Second, in regard to the study of
the relations between the vocalisation and the cantillation, any-
one who wishes to examine the history of the crystallisation of
the vocalisation and cantillation systems as we know them must
also address the phenomenon of the dissonance that is sometimes

found between this vocalisation and this cantillation.

References

Ben-David, I. 1995. Contextual and Pausal Forms in Biblical He-
brew: Syntax and Accentuation. Jerusalem: Magnes Press.
[Hebrew]

Ben-Yehuda, Eliezer. 1908-1958. The Complete Dictionary of An-
cient and Modern Hebrew. Jerusalem: Langenscheidt. [He-
brew]

Biesenthal, Johann H., and F. Lebrecht (eds). 1847. Sefer ha-Shor-
ashim le-Rabi David ben Yosef Qimhi ha-Sfaradi, ‘im ha-Ni-
mugqim me-Rabi ’Eliyhu ha-Levi, ha-Ashkenazi (Rabbi Davidis
Kimhi Radicum liber sive Hebraeum bibliorum lexicon:
Xum animadversionibus Eliae Levitae). Jerusalem: no pub-

lisher. [Hebrew; Latin]

286 Breuer

Blau, Joshua. 1981. ‘On Pausal Lengthening, Pausal Stress Shift,
Philippi’s Law and Rule Ordering in Biblical Hebrew’. He-
brew Annual Review 5:1-13.

Breuer, Mordekhai. 1980. ‘Le-veruran shel Sugyot be-ta‘ame ha-
Miqra u-v-Niqudo, 3: Hilufe ha-Niqud ben Ben-Asher le-
Ven Naftali’. Leshonenu 44: 243-62.

. 1981. ‘KETIV, QERE AND Té'IM (written, read and

chanted)’. Leshonenu 41: 260-69.

. 1982. Ta‘ame ha-Miqra be-Khaf ’Alef Sefarim u-ve-Sifre

Emet. Jerusalem: Mikhlala.

. 1989. Torah, Nevi?im, Ketuvim ‘al pi ha-nusah ve-ha-meso-

rah shel Keter "Aram Tsovah ye-khitve yad ha-qrovim lo, b-ide
Mordekhai Breuer. Jerusalem: Mossad Harav Kook.

Eichel, I. A. 1790. Sefer Mishle ‘im Targum ’Ashkenazi u-Ve?ur. Ber-
lin: Chevrat Ne‘arim.

Finkelstein, Louis (ed.). 1983. Sifra on Leviticus: According to Vat-
ican Manuscript Assemani 66. New York: Jewish Theological
Seminary of America.

Ginsburg, Christian David. 1880-1885. The Massorah. London:

Fromme.

. 1926a. The Later Prophets. London: British and Foreign
Bible Society.

. 1926b. The Writings. London: British and Foreign Bible
Society.

. 1966. Introduction to the Massoretico-critical edition of the

Hebrew Bible. With a prolegomenon by Harry M. Orlinsky.
New York: Ktav.

Dissonance between Masoretic Accentuation and Cantillation 287

Hakham, Amos. 1970. Sefer Iyyov. Da‘at Miqra. Jerusalem: Mos-
sad Ha-Rav Kook.

. 1981. Sefer Tehillim. Da‘at Miqra. Jerusalem: Mossad Ha-

Rav Kook.

. 1984. Sefer Yesha‘yahu. Da‘at Miqra. Jerusalem: Mossad

Harav Kook.

Lipschiitz, Lazar (ed.). 1965. Mishael ben Uzziel’s Treatise on the
Differences between Ben Asher and Ben Naphtali. Jerusalem:
Magnes. [Hebrew]

Norzi, Y. S. 1742-1744. Minkhat Shai. Mantua: Refael Chayim.

Qafih, Y. 1973. Iyyov im Tirgum u-Ferush ha-Ga’on Sa‘adiya ben
Yosef Piyumi. Jerusalem: Hamaqor.

Schlesinger, A. 1962. Researches in the Exegesis and Language of
the Bible. Jerusalem: Ha-Hevra le-Heger ha-Miqra_be-
Yisra’el. [Hebrew]

Sperber, Alexander (ed.). 1959-1973. The Bible in Aramaic. 5
vols. Leiden: Brill.

Wickes, William. 1887. A Treatise on the Accentuation of the
Twenty-One So-Called Prose Books of the Old Testament. Ox-
ford: Clarendon Press.

Yalon, Hanoch. 1971. Studies in the Hebrew Language. Jerusalem:
Bialik. [Hebrew]

Yeivin, Israel. 1968. The Aleppo Codex. Jerusalem: Magnes Press.
[Hebrew]

. 1980. Introduction to the Tiberian Masorah. Translated and

edited by E. J. Revell. Masoretic Studies 5. Missoula, MT:

Scholars Press, for the Society of Biblical Literature and the

International Organization for Masoretic Studies.

288 Breuer

Zer-Kavod, A. 1973. Qohelet. Da‘at Miqra. Jerusalem: Mossad Ha-
Rav Kook.

WHY ARE THERE TWO SYSTEMS OF
TIBERIAN TE‘AMIM?

Daniel J. Crowther

Unlike their Palestinian and Babylonian cousins, the Tiberian
te‘amim' are found in two self-contained systems: one system for
Psalms, Proverbs, and most of Job, and another system for the
other books of the Hebrew Bible.” By convention, the te‘amim
unique to Psalms, Proverbs, and Job are referred to as the te‘amim

of the Three and the te‘amim used in the rest of the Hebrew Bible

‘In this paper, the Hebrew term o’nyv is used to denote the diacritical
marks that are variously referred to in the literature as ‘cantillation
marks’, ‘biblical accents’, and ‘masoretic punctuation’, alongside many
other variations of these names. The transliterated Hebrew te‘amim is
preferred as it suggestive of all the functions of these marks. The English
terms refer to more specific and limited functions, such as the marking
of stress (accentuation), chant (cantillation), and syntax (punctuation),
all of which can be included in a broad understanding of ‘the sense
(ta‘am) of the text’ (Jacobson 2002, 3-24).

? There are many different forms of Palestinian and many different
forms of Babylonian te‘amim found in the manuscripts (and fragments
of manuscripts). These texts bear witness not only to two traditions (be-
side the Tiberian) of marks for the te‘amim, but also to a process of
development of the technology of te‘amim (Heijmans 2013; Shoshany
2013).

290 Crowther

are referred to as the te‘amim of the Twenty-One.* This paper ad-

dresses three questions:

Why are there two kinds of Tiberian te‘amim?

2. What are the features of the books of Psalms, Proverbs,
and (most of) Job that might best explain why these three
books alone have been selected for a different system of
te‘amim?

3. What is the essential feature of the system of the te‘amim
in the Three that distinguishes it from the system of the
te‘amim in the Twenty-One?

1.0. The Absence of Answers in Masoretic

Treatises

Extant masoretic treatises that refer to the te‘amim are few in
number and limited in the degree to which they can be used to
answer any of the above questions. In so far as they do address
the te‘amim, these works appear to focus on the oral performance
of the biblical text and so the function and workings of the

te‘amim tend to be presumed rather than explained. In regard to

° Following the tradition of counting the total number of books in the
Hebrew Bible as 24, as exemplified in, for example, b. Bava Batra 14b.

4 The world of early masoretic treatises is a unique place of particular
interests. Much of the discussion focuses on the observation of minor
grammatical variations in Hebrew that prove rules also observed in Ar-
abic linguistics. A second significant focus appears to have been the
preservation of the correct pronunciation of the biblical text with its
Tiberian markings, whether or not it seemed to follow these rules (Khan
2000, 5-25; 2013).

Why Are There Two Systems of Tiberian Te‘amim? 291

the te‘amim, the two most important masoretic treatises are the
tenth-century Sefer Diqduqe ha-Te‘amim (‘The Book of the Fine
Details of the Te‘amim’, in Hebrew) by Aaron ben Asher; and the
eleventh-century Hiddayat al-Qari’ (‘Guidance for the Reader’, in

Arabic) by the Karaite grammarian Abt al-Faraj Harun.

1.1. Sefer Diqduge ha-Te‘amim

The earliest copy of this work is found in the appendices of the
Leningrad Codex, the opening folio of which dates the comple-
tion of the whole codex to 1008/9 CE.° In the centre of an ornate
and colourful end-folio (479r), the scribe reveals himself to be
Samuel ben Jacob and declares himself to be a student of Aaron
ben Moshe ben Asher (the author of Sefer Diqgduge ha-Te‘amim).
There is no reason to doubt these declarations—indeed the accu-
racy of the biblical text of the codex rather tends to support it. In
the Leningrad Codex, Samuel ben Jacob’s copy of Sefer Diqduge
ha-Te‘amim extends to ten folios (479v—488r).° The material

found in these ten folios can be summarised as follows:

° The text of Sefer Diqduge ha-Te‘amim appears to be in the same hand
as the rest of the Codex.

© The text of Sefer Digduge ha-Te‘amim is available in facsimile copies of
Codex L. Many later copies are found in other codices, but there is much
variation in their material, some of which is found in Baer and Strack
1970 (1879). Aron Dotan (1963) produced a critical edition from the
many variants. The resultant text is much more concise than that found
in Codex L and so contains even less information pertinent to our ques-
tion.

292 Crowther

1. An introduction that offers various blessings and relates
scripture to the twenty-two letters of the Hebrew alpha-
bet (479v-480r).

2. Two piyyutim (liturgical poems) in which the nequddot’
and te‘amim are described and praised (480r—481v).

3. The main bulk of the work comprises lists of words that
have the same consonants, but varied pronunciation ac-
cording to variations in their nequddot and te‘amim
(481v-488r).

From this evidence it can be concluded that Sefer Diqduge ha-
Te‘amim was not written to explain the workings of the te‘amim.
Information about the form and nature of the te‘amim is, in fact,
limited to that found in its two piyyutim, of which one praises the
beauty and efficacy of the te‘amim of the Twenty-One and the
other praises the beauty and efficacy of the te‘amim of the Three.*®
These two devotional poems include word-plays made out of the
names of the te‘amim. These word-plays describe the form and
function of each ta‘am in a way presumably designed to amuse
the reader already familiar with this form and function. For ex-
ample, the ta‘am of the Twenty-One commonly known as geresh
is called in the first piyyut by the alternative name teres (070),
which is likened in form to a net (079 peres) or a hook (07) geres),

” This is the way the naqdan of this text (presumably Samuel ben Jacob)
refers to what is often, by convention, referred to as niqqud by many
scholars today.

8 For a translation of these piyyutim and a discussion of their contents
see Crowther (2015, 48-65).

Why Are There Two Systems of Tiberian Te‘amim? 293

and these can be connected to one another without any destruc-
tion (077 heres). To enjoy this riddle one must know that geresh
is written as a supra-linear angled line (a hook), that the symbol
has a variant, gershayim, which is written as two lines (a net) and
that two geresh signs, or even a geresh and gershayim, are allowed
by the grammar of the te‘amim to follow one another without
either of them being transformed (that is, without 077, destruc-
tion). All of this is most helpful and quite entertaining to the ini-
tiated, especially when the piyyut is vocalised by the person read-
ing it, for it is in this way (and perhaps this way alone) that it
connects the oral to the visual in an oral mnemonic. These oral-
visual mnemonics are, however, rather lost on the uninitiated
and of little, if any, use in understanding why it might be that
there are two systems of Tiberian te‘amim.

This summary begs the question as to why this treatise
should be entitled Sefer Diqgduge ha-Te‘amim when so much of its
contents concern the nequddot. The answer to this question ap-
pears to be that the te‘amim are understood in masoretic writings
to be the determinants of the nequddot (and not vice versa).
Therefore, it is the te‘amim that are understood to be the ultimate

embodiment of the wonders of the oral reading tradition.

1.2. Hiddyat al-Qarv’

Many different parts of this Judaeo-Arabic treatise were trans-

lated into Hebrew in the medieval period in order to be added to

294 Crowther

the appendices of various codices of the Hebrew Bible.’ The focus
of the treatise is the correct (Tiberian) pronunciation of letters
and vowels, for example, in words with ta‘am mille‘el (penulti-
mate stress) and ta‘am millera‘ (ultimate stress). The translation
of the section relating these issues to the te‘amim was variously
copied and came to be known as a treatise in its own right: Sefer
Ta‘ame ha-Miqra. The paragraphs that directly concern the
te‘camim, however, are limited. These paragraphs simply list the
names of twelve te‘amim of the Twenty-One (here understood in
the sense of disjunctive te‘amim)'® and their eight ‘servants’
(which we would call conjunctive te‘amim).'! The discussion of
the te‘amim concludes with a list of which ‘servant’ (conjunctive
ta‘am) is associated with which (disjunctive) ta‘am.'? Wickes
(1881, 104, n. 11) concludes that the name Sefer Ta‘ame ha-Miqra

° Translations of select parts from Hiddyat al-Qadri’? have been copied
from various codices and published under a number of names: for ex-
ample, Sefer Ta‘ame ha-Migra (Mercerus 1565, repr. 1978); Manuel du
lecteur, d’un auteur inconnu publié d’aprés un manuscrit venu du Yémen et
accompagné de notes (Derenbourg 1871). For the first convincing argu-
ment that Hidayat al-Qari’ is the source of these works, see Eldar (1992,
33-42). Regarding the relevance of this work to the te‘amim, see Eldar
and Ofer (2018).

© In the Derenbourg manuscript these names are highly recognisable:
pazer qaton, telisha gedola, teres (geresh), yetiv, zaqef, °atnahta, zarqa, le-
garmeh, tevir, revia‘, tifha, and silluq (1871, 72).
! Shofar munah, telisha getanna, ’azla, merkha, darga, mayela, and galgal.
2 The discussion in Mercerus (1978, 38-44) focuses on five te‘amim:
namely, zargqa, legarmeh, revia‘, tevir, and sillug.

Why Are There Two Systems of Tiberian Te‘amim? 295

is “a misnomer, for the greater part of the work is not taken up
with the ony, but with the op.”

The reproduction of Sefer Diqduqge ha-Te‘amim and Hiddayat
al-Qari’ in the appendices of biblical codices testifies to both to
the high regard in which the te‘amim were held by the medieval
Jewish community and the scarcity of masoretic source texts that
describe their function. These two treatises present a simple hi-
erarchical understanding of the grammar of the te‘amim, in which
disjunctive te‘amim are understood to be emperors, kings, dukes,
and lords, each served by an appropriate conjunctive servant.
Neither of these treatises, nor any other extant masoretic treatise,
makes any attempt to explain the rationale behind the Tiberian
te‘amim, that is, whether they are primarily accents, or primarily
punctuation marks, or primarily cantillation marks (or all of
these or something more), let alone why there should be two sys-

tems of Tiberian te‘amim.

2.0. A First Answer: Poetics

According to James Kugel (1981, 109-16), during the late medi-
eval and early-modern period, the te‘amim of the Three were
widely identified as representing the essence of Biblical Hebrew
poetry. This, according to Kugel, led to the “forgetting of paral-
lelism” as the mainstay of biblical poetics. Thus Job 1.1-3.1 and
42.7-17 came to be understood to have been marked with the
te‘amim of the Twenty-One because they were ‘essentially’ prose
in form, whilst Job 3.2-42.6 were understood to have been
marked with the te‘amim of the Three because these chapters

were poetic in form. There are, however, good reasons to doubt

296 Crowther

that this approach was ever predominant in rabbinic thought,
even if it did successfully infiltrate much of the thinking of early-
modern Christian Hebraists.

First, the understanding of the te‘amim of the Three as es-
sentially poetic and of the Twenty-One as essentially prosaic is
not found in any masoretic treatises. The rabbinic readers of
these treatises, therefore, should have been aware that all the ex-
tant masoretic treatises and piyyutim declare both systems of
te‘amim to give (poetically) enlightened performances of the texts
in which they are found.

Second, it is hard to imagine that any Jewish rabbi could
have been unaware either that there are important poetic texts
in the Torah (notably Gen. 49, Exod. 15, and Deut. 32) or that
these texts are presented with the te‘amim of the Twenty-One.

Third, the Decalogue is presented in masoretic codices with
two sets of te‘amim: one for the high (poetic?) cantillation of the
trained reader on feast-day and one for the low cantillation of
household readings (Cohen and Freedman 1974, 7-19). The high
cantillation is presented with the te‘amim of the Twenty-One, not
the Three.

Fourth, there are many poetic texts outside the Torah pre-
sented with the te‘amim of the Twenty-One (notably Judg. 5, 2
Sam. 22, and 1 Chron. 16). Any rabbi who proposed that Job
switches between the two systems for reasons of poetry and prose
would have also needed to provide some cogent explanation of
why other books of the Hebrew Bible do not also switch systems
of te‘amim when their texts also switch from prose to poetry (and

back again).

Why Are There Two Systems of Tiberian Te‘amim? 297

Fifth, the title Song of Songs (shir ha-shirim) declares it to
be a poetic work, as does one of the rabbinic titles of the book of
Lamentations (qinot). If it is held that the te‘amim of the Three
are the poetic te‘amim, it is far from clear why these two books
have been given poetic titles when they are presented with the
te‘amim of the Twenty-One.

Sixth, consistent with all the above, the authoritative six-
teenth-century work Masoret ha-Masora by Elijah Levita (Eliyahu
Bahur ha-Levi) praises the poetical virtues of both systems of
te‘amim, not just the te‘amim of the Three and does not refer to
the te‘amim of the Three as being essentially ‘poetic’ (Elijah Levita
in Ginsburg [ed.] 1867).'*

3.0. A Second Answer: Verse Length

The titles of William Wickes’s two works on the te‘amim refer to
the “so-called poetic books” (1881) and the “so-called prose
books” (1887), both acknowledging the terminology that by the
nineteenth century had become conventional and casting doubt
upon it.'‘* To counter this misnomer, Wickes refers to the elev-
enth-century writings of Rabbi Yehuda ben Bil‘am (Sephardi) and
the twelfth-century tosafot of Rabbi Isaac ben Meir (ben
Yokheved bat Rashi). Both these works understand verse length,

'3 According to Ginsburg, all his other works are similarly descriptive
rather than explicative, including Levita (1538). Nevertheless, in his in-
troduction to the work, Ginsburg (1867, 65, n. 71) himself does refer to
the te‘amim of the Three as the ‘poetic accents’.

Cf. also Davidson’s (1861) nineteenth-century introductory work
Outlines of Hebrew Accentuation, Prose and Poetical.

298 Crowther

not poetics, to be the distinguishing mark of the te‘amim of the
Three. According to Wickes (1881, 8-9): “The idea seems to have
been to compensate for the shortness of the verses (which is the
marked characteristic of the greater part of these books) by a
finer and fuller, more artificial and impressive melody.”

Table 1: The average number of words per verse in each book of the
Hebrew Bible

Book Word Verse Words/ Book Word Verse Words/

Total Total Verse Total Total Verse
Prov. 7034 915 7.7 Amos 2060 146 14.1
Ps. 19642 2527 7.8 Jon. 690 48 14.4
Job 8428 1070 7.9 Zeph. 774 53 14.6

Lam. 1650 154 10.7 Ezek. 19033 1273 15.0
Song 1270 117 (10.9 Zech. 3166 211 15.0
1 Chron. 10962 943 11.6 Deut. 14465 959 15.1

Nah. 565 47 12.0 Josh. 10035 658 15.3
Hab. 677 56 12.1 Ruth 1303 85 15.3
Hos. 2391 197 12.1 Hag. 607 38 16.0
Num. 16540 1289 12.8 Mal. 883 55 16.1
Neh. 5428 405 13.4 Judg. 9922 618 16.1
Mic. 1411 105 = 13.4 2Sam. 11206 695 16.1
Joel 964 73 13.2 1Kgs 13234 817 16.2
Isa. 17157 1291 13.3 Jer. 22230 1364 16.3

Gen. 20632 1533 13.5 2 Chron. 13474 822 16.4
Eccl. 3000 222 13.5 1Sam. 13447 811 16.6
Obad. 292 21 13.9 Dan. 6054 357 17.0
Exod. _ 16880 1213 13.9 2Kgs 12389 719 = 17.2
Ezra 3911 280 14.0 Est. 3078 167 18.4
Lev. 12058 859 14.0

Three books of the Hebrew Bible average less than eight
words per verse: Psalms (7.7), Proverbs (7.8), and Job (7.9). All
other books have average verse lengths of twelve words or more,

Why Are There Two Systems of Tiberian Te‘amim? 299

except Lamentations (10.7), Song of Songs (10.9) and 1 Chroni-
cles (11.6).!° The latter three contain significant sections of short-
verse text alongside other sections with long verses. Table 1,
therefore, shows that there are two kinds of books in the Hebrew
Bible: short-verse books (with less than eight words per verse)
and long-verse books (with more than twelve words per verse).
According to Wickes, the te‘amim of the Three were a practical
Tiberian response to the challenge of punctuating (or cantillat-
ing) two different kinds of text: short verse texts as against long
verse texts.

The credibility of the long-verse/short-verse explanation
rests on two foundations. First, whether or not verse lengths were
determined prior to the time of the Tiberian Masoretes—only if
they were, could verse length have led to the creation of two sys-
tems of te‘amim (and not vice versa). Second, how one also might
explain cases like Lamentations, Song of Songs, and 1 Chronicles:
that is, texts that contain clearly defined sections of short verse

material that is presented with the te‘amim of the Twenty-One.

3.1. The Priority of the Tradition of Verse Division

For Wickes, the existence of an established tradition of verse di-
vision prior to the Tiberian Masoretes was suggested by the men-

tion of pesuge ha-tora ‘verses of the Torah’ and pesuge ha-te‘amim

'S 1 Chron. and 2 Chron are presented in this table as two books to
highlight the variations of verse length within the two halves of one
work (Chronicles). The same logic was extended to Samuel and Kings.
2 Sam. contains two lengthy poems, without which its average verse
length would have exceeded that of 1 Sam.

300 Crowther

‘verses of the te‘amim’ in the Mishna (Megillah 4.4.) and Talmud
(b. Berakot 19a; 62a—b; b. Qiddushin 32b; b. Yoma 52a-b; b. Ne-
darim 37b). Whilst it is far from certain that the verse divisions
mentioned in the Mishna and the Talmud are identical to those
of masoretic tradition (Blau 1896; 1897), a number of texts in the
Dead Sea Scrolls do have spaces that indicate traditions of para-
graph, section, and even verse division. The first-century CE
Isaiah Scroll (1QIsa*), for example, has many paragraph and sec-
tion divisions that are consistent with the later masoretic tradi-
tion and the first-century CE Great Psalm Scroll (11QPs*) has end-
verse spaces which are entirely consistent with the later Maso-
retic traditions for Ps.119 and 145 (but less so for other psalms)
(Burrows, Trever, and Brownlee 1965). Furthermore, the many
hundreds of biblical texts with Palestinian and Babylonian
tecamim recovered from the Ben Ezra Synagogue, Fustat, Old
Cairo (‘the Cairo Genizah’) display verse divisions (nearly) iden-
tical to those found in the later standard (Tiberian) Masoretic
Text (Kahle 1927; 1966; Revell 1970, 157-99). In other words,
whilst the tradition of verse divisions for many books of the He-
brew Bible may not have been finalised until after the time of the
Mishna, and perhaps not even until after the period of the Tal-
mud, we can be confident that the Tiberian Masoretes were the
recipients, not the creators, of the versification of the biblical

text.

3.2. Lamentations

In BHS, all the verses of the first two chapters of Lamentations

are presented as three poetic lines (except 1.7 and 2.19, which

Why Are There Two Systems of Tiberian Te‘amim? 301

are presented as four) (Elliger et al. [eds.] 1977, 1354-67).
Whilst the division of each of these verses into stichs is a matter
of some debate, it is indisputable that all of these verses are long.
They vary between fourteen and twenty words in length. The
verses in ch. 4 are not so long. In BHS they are presented as two
poetic lines (typically with four stichs) and they vary between
eight and seventeen words in length.’° The verses in chs 3 and 5
are, in comparison, shorter. All of the 66 verses in ch. 3 have
fewer than eight words (bar one verse);’’ and all bar two of the
22 verses in ch. 5 have fewer than nine words.'* Unlike chs 1, 2,
and 4, the number of poetic lines found in these verses is indis-
putable: they must be read as single poetic lines, typically with
two stichs each.

If the system of te‘camim employed had been determined by
line length alone, one might expect the te‘amim of Lamentations
to switch between the two Tiberian systems, just as in Job. In the
case of Job, however, the switch from the te‘amim of the Twenty-
One to the te‘amim of the Three at Job 3.2, and then back at 42.7,
follows not only verse length, but also the literary style and con-
tent of the text. This switch of te‘amim in Job, therefore, presents
the book of Job with an introduction, a central discourse, and a
conclusion. In the case of Lamentations, the switches between
long verse length (chs 1 and 2), short verse length (ch. 3), me-
dium verse length (ch. 4), and short verse length (ch. 5) do not

17 Lam. 3.22 has nine word-units.

18 Tam. 5.1 and 5.17 have nine word-units.

302 Crowther

mark any changes in literary content or style (other than line
length). A switch between the systems of te‘amim in Lamentations
may have thus undermined its unity of style and content.

The use of the te‘amim of the Twenty-One for long verses
and short verses in Lamentations indicates that the te‘amim of the
Twenty-One can be used for short verses when they are called
upon so to do. There are, however, some problems generated
when the te‘amim of the Twenty-One are so employed—an issue

to which we will return.

3.3. Song of Songs

Song of Songs is presented with 24 short verses and 83 long
verses. Unlike Lamentations, these short verses are interspersed
between the long verses. The argument for the exclusive use of
one form of te‘amim (by default, that of the Twenty-One) is thus
even more compelling. The use of the te‘amim of the Twenty-One
for the short verses of Song of Songs again confirms that the
tecamim of the Twenty-One can be used for short verses when

they are called upon so to do.

3.4. 1 Chronicles

The low average verse length of the half-book 1 Chronicles (the
sixth lowest, with 11.6 words per verse) is at odds with that of
its other half, 2 Chronicles, which has the fifth highest (16.4).
This anomaly is explained by the many short-verse genealogical
lists that dominate 13 of the 29 chapters in 1 Chronicles (1 Chron.
1.1-9.44; 11.26-47; 12.24-37; 24.7-18; and 25.9-31). 1 Chron.

16 also contains an extensive quotation of short-verse material

Why Are There Two Systems of Tiberian Te‘amim? 303

that is parallel (near identical) to the text of three psalms.'° Since
both the genealogical and the Psalm material is distinct in style
and content from the surrounding narrative, presumably this ma-
terial could have been presented with the te‘amim of the Three.
The fact that it is not suggests that there was more involved in
the decision to employ the te‘amim of the Three than verse length.

In the genealogical lists of 1 Chronicles, many of the verses
have five or fewer words and many others are short verses of
eight or fewer words. In most of these verses, an ’atnah is found
at the point of its most significant semantic division. The gram-
mar (rules) of the te‘amim of the Twenty-One require a disjunc-
tive tifha to precede an ’atnah on one of the two word-units before
the word with ’atnah (Wickes 1887, 69; Price 1990, 58-61).7° A
similar rule requires a tifha to occur on one of the two words
before a silluq (Wickes 1887, 62; Price 1990, 54-57). As a result
of these two rules, four disjunctive te‘amim (that is, tifha—’atnah—
tifha-silluq) must be used in every verse presented with ’atnah,
even if the verse itself contains a total of only five, six, or seven
words. Consequently, on verses of fewer than eight words, the
rules of the te‘amim of the Twenty-One produce a most predicta-
ble pattern of recitation. When these short verses occur consecu-

tively, for example in short-verse poetry or in genealogical lists,

1° 1 Chron. 16.8-36. The material follows Ps. 105.1-15; 95.1-13; and
106.1, 47-48. The dependency of 1 Chronicles on the Psalter is here
presumed for the sake of simplicity. The discussion here regarding the
use of different kinds of te‘amim is not affected by this presumption.

° This is a general rule with the exceptions of mayela (16 cases) and
when both words preceding ’atnah are monosyllabic (31 cases). See
Yeivin and Revell 1979, 177-81; Jacobson 2002, 69-71.

304 Crowther

this predictability will cause the recitation to sound repetitive. In
the case of many literary genres, such as the psalm material in 1
Chron. 16.8-36, a repetitive recitation may be considered prob-
lematic. In the case of the genealogical lists, such as those found
in 1 Chron. 1.1-9.44, a predictable and repetitive recitation may
be considered most appropriate. It is striking, therefore, that
whilst the short-verse poetic material quoted in 1 Chron. 16.8-
36 does not use any ’atnah te‘amim to delimit the parallel stichs,
most of the short-verse material in 1 Chron. 1.1-9.44 does.

In the Psalter itself, however, all the material quoted in 1
Chron. 16.8-36 does employ ’atnah te‘amim. This does not result
in a repetitive recitation, because under the rules of the Three,
an ’atnah does not require a disjunctive ta‘am to precede it. Fur-
thermore, the rules of the Three forbid the occurrence of a dis-
junctive ta‘am too close to the ’atnah or silluq; that is, two full
syllables must separate the syllable with ’atnah or silluq and the
syllable with the preceding disjunctive ta‘am (Wickes 1887, 60,
69, 75; Price 1990, 209-13, 234-38).

In 1 Chron. 16, when the Psalms material is quoted, zagef
te‘camim are employed where ’atnah te‘amim are found in the Psal-
ter. This is because zagef te‘amim in the Twenty-One, like ’atnah
tecamim in the Three, do not require a preceding disjunctive
ta‘am. This observation has led some commentators to declare

that an ’atnah in the Three is equivalent (in pausal effect) to a

Why Are There Two Systems of Tiberian Te‘amim? 305

zagef in the Twenty-One. Before rushing to this imperfect conclu-
sion,”! however, it would be wise to cast our net a little further
and examine, first, the presentation of other poetic texts in the
Twenty-One and, second, other texts in the Three that are paral-

leled by texts in the Twenty-One.

3.5. Short-Verse Poetic Texts with the Te‘amim of the
Twenty-One

All the short verses of Lamentations chs 3 and 5 are presented as
poetic lines with two stichs (that is, as parallelism). In ch. 5, these
stichs are delimited by zagef te‘amim: (and ’atnah is not used in
any of the verses). In ch. 3, the delimitation of the stichs is a little
more complex: zagef te‘amim are used in forty-seven of its sixty-
six verses; ’atnah is used once (Lam. 3.56, for a division so unu-
sual, it is not recognised by the stichography of BHK, BHS, or
BHQ; see Kittel et al. [eds.] 1937, 1238; Elliger et al. [eds.], 1977,
1367; Jan de Waard 2004); in the remaining eighteen verses tifha
te‘amim delimit the first stich.” The use of tifha to delimit the first
stich is interesting. It allows the second stich (the stich delimited
by silluq) to be without more smoothly. If zagef tecamim had been

used in these verses, then a tifha would have been required before

71 It is clear that this conclusion is imperfect since zagef te‘amim do not
prohibit a disjunctive ta‘am occurring closer than two full syllables be-
fore them whilst the ’atnah of the Three does. In other words, whilst
zagef te‘amim in the Twenty-One can perform in a similar way to ’atnah
te‘amim in the Three, their identity and function is defined by the system
of the Twenty-One and this system is different to that of the Three.

22 Vv. 2,3, 5, 7,11, 15, 17, 19, 21, 30, 31(?), 45, 46, 47, 49, 52, 54, 58,
and 64.

306 Crowther

the silluq (by the rules of the Twenty-One) and this would have
added a (presumably unwanted) “pause” to the second stich.”

According to Price (1990, 72), the reason why zagef
te‘amim are used to delimit the stich of poetic texts in the Twenty-
One is because “the domain of Little Zagef is the most complex
and flexible of all the other accents.” In practice what this means
is that zagef can repeat to give one, two, three, or four stichs and
can occur without a preceding lesser disjunctive ta‘am. This flex-
ibility is not unlimited: if there are more than two word-units in
the clause of zagef, either a pashta or yetiv disjunctive is required.
Even here, however, the judicial use of maqgefim can ensure that
the “pauses” in each stich are determined solely by poetics and
not the grammar of the te‘amim.**

The book of Lamentations can thus be understood as
providing the paradigm for two kinds of presentation for poetic
texts of the Twenty-One:

?3 As the proponents of prosody have rightly observed, oral segmenta-
tion is achieved by multiple means including intonation, stress, empha-
sis, and melody. For simplicity, these multiple tools are referenced here
as “pause”. The intended sense of “pause” here is the delimitation mark-
ers of oral segmentation. Whether or not there is any period of cessation
or silence (a literal pause) would, of course, be determined by the read-
ers chosen method of orally performing the text. The relevance of this
to the study of the te ‘amim is particularly clear in Pitcher 2021.

°4 Zagef te‘amim are used without any ’atnah te‘amim in eight of the 110
verses of 1 Chron. 1-2 (1.14, 28, 30; 2.12, 14, 20, 37, 41, 51). It is
interesting to note that when the lists of 1 Chron. 24.7-18 and 25.9-31
do employ zagef te‘amim in this way, the pashta is never lacking. This is
in accord with the poetics of a list. It creates a clear and repetitive rec-
itation.

Why Are There Two Systems of Tiberian Te‘amim? 307

Presentation 1 [P1]: short verses (nine words or less) with
one poetic line of two-stich classic parallelism. In these
short verses, ’atnah te‘amim are absent and either zagef or
tifha te‘amim delimit the first stich.

Presentation 2 [P2]: long verses (nine words or more) that
contain more than two stichs. In these longer multi-stich
verses, more complex forms of parallelism may be observed
and a combination of ’atnah and zagef te‘amim (alongside
other disjunctive te‘amim) will be employed to delimit each
verse into these stichs.”

As previously mentioned, in the Song of Songs (unlike Lamenta-
tions), long and short verses are interspersed amongst its eight
chapters. In 23 of the 25 verses with eight or fewer word-units,
the accentuation follows the first presentation above [P1],”°
whilst all those of nine word-units or more follow the second
[P2]. The two exceptions are Song 2.5 and 4.15. These are excep-
tions that prove the rule. For whilst these two verses do both have
eight word-units, both can be read (with the stichography of
BHK; see Kittel et al. [eds.] 1977, 1202, 1206) as having three
stichs, the first of which is delimited by a zagef and the second
by an ’atnah.

Outside of Lamentations, Song of Songs, and 1 Chronicles,

two-stich short verses of poetry are observed with ’atnah te‘amim

5 This observation is self-evident, but I have not found it explicitly
stated elsewhere.

© The 23 verses are Song 1.1, 2, 1.9-2.2, 4, 6; 3.9; 4.7; 5.10; 6.3; 7.7-
8, and 11. Strictly speaking, the stichs of these verses are delimited by
zagef and silluq te‘amim.

308 Crowther

delimiting their first stichs—that is, with the kind of presentation
observed in the genealogical lists of 1 Chronicles. The crucial ob-
servation here, however, is that these poetic verses do not occur
in a consecutive run of short verses (which would make the reci-
tation repetitive) in all cases bar one. Once again, this text can
be understood to be the exception that proves the rule. The text
in question is the Song of David at 2 Sam. 22, which parallels the
text of Ps. 18 with strikingly similar, but rarely identical, content.
In 2 Sam. 22, 41 of its 51 verses are short verses of eight words
or fewer.”’ All of these short verses are presented as two stichs
and all of these 41 short verses are delimited by ’atnah te‘amim.
The method in 2 Sam. 22 can be taken, therefore, to describe a
third method of presenting poetry with the te‘amim of the
Twenty-One:

Presentation 3 [P3]: short verses (eight words or less) of

two-stich parallelism that employ ’atnah te‘amim to delimit

their first stich.

Table 2 observes the employment of these three presentations for
twelve of the most widely-recognised poetic texts in the Twenty-
One and four prophetic texts also widely considered to be

founded upon the poetics of parallelism.

2? Vv, 1-2 are exceptional, v. 1 forms the title of the poem, v. 2 the
opening ‘and he said’. Vv. 3, 7, and 16 are read here as a quatrains. Vv.
8, 9, 31, 44, 49, and 51 are read as three stichs.

Why Are There Two Systems of Tiberian Te‘amim?

309

Table 2: Poetic Texts in the Twenty-One (see Appendix for details)

Pl P2 % P3 %
Passage Number Short Long __ stichs Short stichs
of verses verses verses delim- verses de- delim-
[stichs] delimited with ited as limited by ited as
by zagef zagef Plor -atnah P3
[stichs] +’atnah P2 [stichs]
[stichs]
Gen. 49.2-27 26 [80] 10) 17 [62] 77% 9 [18] 23%
Exod. 15.1-18 18 [60] 3 [6] 12 [48] 90% 3 [6] 10%
Num. 6.24-26 3 [6] 3 [6] 0 100% 0 0
Deut. 32.1-43 43 [140] 0 26 [106] 75% 17 [34] 25%
Judg. 5.2-30 29 [121] 0 26 [115] 93% 4 [8] 7%
1 Sam. 2.1-10 10 [36] 10) 7 [30] 83% 3 [6] 17%
2 Sam. 1.19-27 9 [40] 1 [2] 8 [38] 100% 0 0
2 Sam. 22.2-50 49 [109] 10) 8 [27] 25% 41 [82] 75%
2 Sam. 23.1-7 7 [23] 10) 6 [21] 91% 1 [2] 9%
Isa. 5.1-7 7 [37] (0) 7 [37] 100% 0 0
Isa. 40.1-31 31 [114] 10) 22 [96] 84% 9 [18] 16%
Hab. 3.2-19 18 [64] 10) 14 [56] 87% 4 [8] 13%
Jon. 2.3-10 8 [28] 10) 7 [26] 93% 1[2] 7%
Lam. 154 [536] 88 [176] 66 [352] >99% 1 [2] <1%
Song 117 24 [48] 87 100% 0 0
1 Chron. 16.8-36 29[61] 25 [50] 4 [11] 100% 10) 0

From Table 2 it can be seen that, apart from Song, Lam. and 1
Chron. 16, the most frequent presentation of the verses of our
sixteen chosen texts is the second presentation type [P2], in
which the poetry is presented in longer verses with multiple
stichs delimited by ’atnah and zagef (alongside other disjunctive
te‘amim) (Renz 2003). In many texts, however, short verses occur

interspersed between long verses, as was observed in Song of

310 Crowther

Songs. Unlike Song of Songs, the other poetic texts in the Twenty-
One commonly employ ’atnah te‘amim to delimit these stichs
[P3]. Since these verses are interspersed within longer verses, the
problem of repetitive recitation does not occur—except in regard
to 2 Sam. 22.

Table 2 highlights the extent to which 2 Sam. 22 is an ex-
ceptional case. Whilst 75 percent of the stichs of 2 Sam. 22 are
presented with the third presentation type [P3], no other poetic
text in the Twenty-One presents more than 25 percent of its stichs
according to this presentation [P3]. Other than 2 Sam. 22, P3
occurs as a minority presentation within the wider context of
longer multi-stich verses [P2].

Outside of 2 Sam. 22, only five of the above poetic texts

have consecutive short verses delimited by ’atnah:

1. Gen. 49.19-21: three verses listing the blessings of Jacob
upon Gad, Asher and Naphtali.

2. Deut. 32.18-19: two verses. Since v. 19 follows from v.
18, the repetition of form is helpful.

3. Deut. 32.33-34: two verses. Since v. 34 responds to v. 33,
the repetition of form is helpful.

4. Isa. 40.16-18: three verses with interlinear parallelism.
Since v. 17 repeats the content of v. 16 and v. 18 responds
to v. 16-17, the repetition of form aids these connections.

5. Isa. 40.29-30: two verses with interlinear parallelism.

It would appear, then, that only Gen. 49.19-21 and 2 Sam. 22.2-
51 are presented so as to be recited in the manner of a list. In the
first case this is not problematic—the material is a list. In the

second, this is puzzling. The Song of David of 2 Sam. 22 is a song

Why Are There Two Systems of Tiberian Te‘amim? 311

of heterogeneous content, passionate emotion, and dramatic sal-
vation. It is not a list, so why should it be presented to be read in
such a repetitive manner?

The treatise Soferim may be of help in answering this ques-
tion. The treatise is a compendium of Talmudic wisdom concern-
ing the correct handling of scriptural scrolls. In this compendium,
two pairs of parallel texts are given special attention: 2 Kgs 18-
20 || Isa. 36-38 and 2 Sam. 22 || Ps. 18 (Cohen 1965).”° These
texts are recognised as being both very similar and yet having
important differences. The differences are thus listed in Soferim
in order to ensure that no scribe—either intentionally or uninten-
tionally—will harmonise the texts of 2 Sam. 22 and Ps. 18. In the
context of this Talmudic wisdom, if the Tiberian Masoretes had
presented 2 Sam. 22 with zagef te‘amim [presentation P1], the
recitation would have had much the same flow of rhythm and
“pause” as Ps. 18. The two texts would have been much more
likely to suffer amalgamation in the mouths, minds, and, there-
fore, hands of their scribal custodians. To avoid this risk, it is
plausible that the Tiberian Masoretes may have elected to present
the short verses of 2 Sam. 22 with ’atnah te‘amim [presentation
P3], even though it has many consecutive short verses. This
presentation generated a repetitive recitation for 2 Sam. 22 dis-

tinct from the recitation of Ps. 18. The extent of its repetitive

8 Soferim (Scribes), Rule One of ch. 8 lists 72 words of 2 Sam. 22, all of
which should be guarded from harmonisation with the different but
similar wording of Ps. 18. The words of Ps. 18 are not listed. Ironically,
some of the vowel letters given as the definitive form of the words of 2
Sam. 22 differ from those found in Codex L.

312 Crowther

nature can be seen in the following comprehensive list of all the

disjunctive te‘amim as they occur in 2 Sam. 22.32-40:

(32) tifha —atnah: tifha — silluq.

(33) tifha —atnah: tifha — silluq.

(34) _ tifha —’atnah: tifha — silluq.

(35) tifha —atnah : tifha — silluq.

(36) tifha —atnah: tifha — silluq.

(37) tifha —?atnah: tifha — silluq.

(38) tifha —atnah: tifha — silluq.

(39) tifha —atnah: tifha — silluq.

(40) tifha —atnah: tifha — silluq.
Anyone reciting (or memorising) 2 Sam. 22 according to these
te‘amim would thus recite vv. 32-40 in a highly repetitive man-
ner. In terms of the dramatic delivery of the varied content of
these verses, this is a bit of a disaster.2? But in terms of textual
memorisation, it would act as a significant reminder that, whilst
2 Sam. 22 is very much the Song of David, it is definitely not the
Song of David of Ps. 18.

The genius of this presentation of 2 Sam. 22 is that whilst

it testifies to the same delimitation of stichs as Ps. 18, it contains
41 additional disjunctive te‘amim. Most of these disjunctive

te‘camim are imposed upon the recitation by its presentation with

-atnah te‘amim [P3], and none impact the delimitation of a stich.

° These verses praise God as a refuge, then as a trainer of my hands for
battle, before recounting the manner of the revenge I have taken on my
enemies.

Why Are There Two Systems of Tiberian Te‘amim? 313

These extra disjunctive te‘amim add many “pauses” and occur in
highly predictable places, which thus stretches the recitation out

to sound more like a repetitive list than a song of varied praise.

4.0. A Third Answer: The Dilemma of Short-verse
Poetry in the Twenty-One

The observed general preference for zagef te‘amim in delimiting
poetry in the Twenty-One is based upon their flexibility. This
flexibility allows a stich delimited by zagef to be read either with-
out any other disjunctive te‘amim (when the poetics so demand)
or with one or more disjunctive te‘amim in any place in the clause
(when this is more appropriate). The flexibility of zagef, however,
comes at a cost. Because it is so flexible, it cannot produce the
same syntactic clarity that is delivered by ’atnah. For this reason,
?atnah te‘amim are preferred as markers of dichotomy in the short
verses that are interspersed throughout Gen. 49 and Deut. 32. In
these texts, the majority presentation is the second [P2], so the
recitation of the whole is not made to sound repetitive by occa-
sional short verses with the third presentation [P3]. The di-
lemma, therefore, of presenting short verse poetic texts with the
te‘amim of the Twenty-One is that whilst the delimitation of their
first stichs by zagef provides flexibility at the expense of clarity,
the delimitation of the first stichs by ’atnah provides clarity at the
expense of flexibility.

A similar dilemma surrounds the use of tifha to delimit the
first stich of a short verse in order that the second stich might be
free from disjunctive te‘amim (see above). Once again, the system

of the Twenty-One is shown to be able to present poetry without

314 Crowther

any unwanted “pauses” generated by the rules of the te‘amim of
the Twenty-One. Again, however, the cost of this flexibility is a
loss of syntactic clarity. The Tiberian solution to this dilemma
was to create a secondary system of te‘amim with similar princi-
ples to those of the Twenty-One, but with more flexible parame-
ters, i.e., a system that employs ’atnah te‘amim that can be pre-
ceded by disjunctive te‘amim, but do not make such a precedent

mandatory.

4.1. A More Flexible Alternative

In the rules of the te‘amim of the Three, the ’atnah is preceded by
a dehi disjunctive ta‘am just as an ’atnah is preceded by a disjunc-
tive tifha ta‘am in the Twenty-One. Unlike in the Twenty-One,
however, a dehi ta‘am can be placed anywhere in a colon delim-
ited by ’atnah (not just on one of the two words preceding the
-atnah) and it need not appear at all.

A similar situation describes the stich delimited by sillug. A
sillug in the Three is preceded by a disjunctive revia‘ mugrash, just
as a silluq is preceded by a disjunctive tifha in the Twenty-One.
Unlike in the Twenty-One, however, a revia‘ mugrash can be
placed anywhere in the colon delimited by silluq (not just on one
of the two words preceding the silluq) and it need not appear at
all.

Wickes (1881, 99-101) expressed his understanding of this
situation in terms of various “laws of transformation.” These laws
explained why his principle of continuous dichotomy was not fol-

lowed in so many verses of the Three. According to these laws of

Why Are There Two Systems of Tiberian Te‘amim? 315

transformation, a disjunctive dehi is transformed into a conjunc-
tive munah whenever it occurred on the word preceding a word
with ’atnah and the stress of its word was not separated from the
stress of the word with ’atnah by two or more ‘full vowel’ sylla-
bles. A similar law explained the absence of an expected revia‘
mugrash in a colon delimited by sillug. Whilst these laws have
much explicative value, the cases of transformation of dehi
tecamim are far more extensive than that delivered by Wickes’s
simple statement of the laws of transformation (see also Price
1990, 36, 209-13, 234-38).

4.2. Ps. 18 as an Example

Dehi te‘amim are absent in Ps. 18 in 26 (53 percent) of the 49
stichs that are delimited by ’atnah: 5a, 6a, 7c, 8b, 9b, 12b, 13b,
16c, 18a, 19a, 20a, 21a, 23a, 24a, 25a, 26a, 27a, 31b, 36b, 37a,
41a, 42a, 43a, 46a, 48a, and 50a. In 17 (65 percent) of these 26
cases, the absence is well explained by the law of transformation:
namely 6a, 7c, 8b, 9b, 12b, 13b, 16c, 21a, 23a, 24a, 25a, 31b,
37a, 41a, 46a, 48a, and 50a.*° In these cases, the syntax places

3° Somewhat obscurely Price (1996, V:1196) lists verses with virtual
dehi as 1, 6, 7, 8, 9, 12, 18, 21, 23, 24, 25, 37, 41, 43, 48, and 51. His
list agrees with the above in regard to twelve verses (6, 7, 8, 9, 12, 21,
23, 24, 25, 37, 41, and 48) and is divergent in regard to four verses (1,
18, 43, and 51): v. 1 has a dehi before the ’atnah, there is a long run of
conjunctives before the revia‘ mugrash, but the reason for a transfor-
mation is not well explained by the laws of transformation; vv. 18 and
43 are considered above; v. 51 has six disjunctive te‘amim on nine word-
units, so it is not clear how an additional dehi disjunctive ta‘am can be
considered to be virtually present (and transformed).

316 Crowther

the position of a preceding dehi too near to the stress of the word
with ’atnah.

Nine cases, however, are not well explained by the ‘Laws
of Transformation,’ specifically, 5a, 18a, 19a, 20a, 26a, 27a, 36b,
42a, and 43a. Three of these nine cases could be explained if the
vowel hireq in a yiqtol verb is not considered to constitute a ‘full
vowel’ syllable (26a, 27a, and 36b), but this proposal generates
more problems than it solves; for survival of the dehi at other
places (for example, Ps. 18.17a and elsewhere in the Psalter)
would then require further explanation. Four further cases (5a,
19a, 42a, and 43a) could be explained if a dehi ta‘am is consid-
ered to be transformed when the word with ’atnah has maqgefim.
But, again, the survival of dehi in other psalms under these con-
ditions would then require explanation (for example, at Ps.
22.18a). At Ps. 18.18a, it is not clear that a dehi should neces-
sarily be expected on "28 ‘from my enemy’ before the short
word ty ‘strong’. The two words are usually read together to mean
‘from my strong enemy’. Although it must be admitted that a def-
inite article would be expected on tv. Furthermore, ‘you save me,
my enemy is strong’ does form a nice parallel with 18b. The case
of 20a is clearer. The ’atnah occurs on the last syllable of an7/3?
‘to the broad place’. This word is long enough to protect a dehi
on the preceding word from transformation (7ix°¥i1 ‘and he
brought me out’), but no dehi is found upon it.

A similar situation pertains to the revia‘ mugrash, which is
absent in thirteen (25 percent) of the 51 stichs delimited by silluq:
2, 5b, 15b, 19b, 28b, 33b, 36c, 40b, 43b, 44c, 48b, 50b, 51c. In

eight of these thirteen cases the law of transformation explains

Why Are There Two Systems of Tiberian Te‘amim? 317

the absence of revia‘ mugrash: 5b, 28b, 33b, 36c, 40b, 43b, 48b,
and 50b. Three of the remaining five cases have other explana-
tions. The stichs of vv. 2 and 51c may be considered grammati-
cally exceptional. At 15b there is a mugrash symbol without a
revia‘.. This probably should be read as an alternative representa-
tion of revia‘ mugrash: the anomaly occurs elsewhere, it is con-
sistently replicated in many of the best masoretic manuscripts,
and it occurs regularly on monosyllabic words (Dotan 2001, xvi).
In regard to 19b, however, one is required to argue, against sense,
that the syntactic or prosodic dichotomy is expected between
www and *) as a support for me’ (and not after 717 ‘the LoRD’).
The absence of revia‘ mugrash at 44c cannot be explained by the
laws of transformation at least in their current guise. Despite
these observations, it is not argued here that Wickes’ laws of
transformation cannot be stretched to accommodate these and
multiple similar cases.*! Rather, it is asked whether there is good
purpose to this exercise. What is achieved by these ever-more

complex explanations is the preservation of a system of rules of

3! Consider, for example, the cases of the transformation of dehi in the
first four ssalms of the Psalter. In Ps. 1.3; 2.2; and 3.9 more than two
syllables separate the stress of the word with ’atnah and the stress of the
preceding word. The transformed ta‘am preceding the °atnah in these
cases is merkha (not munah, as per Wickes’s Law of Transformation). In
Ps. 2.7 and 4.9, the expected positions of their syntactic dichotomies
occur on the second word preceding the ’atnah and not the preceding
word. In both cases these words also have merkha conjunctive te‘amim.
Further rules are needed to explain these transformations. For various
lists of virtual dehi te‘amim see Price (1996, V:1195-210).

318 Crowther

the te‘amim that has been supposed to explain them. What is frus-
trated is an observation of the recitation to which the te‘amim
bear witness (which may in itself provide good explanation for
these te‘amim). It seems to me intuitively sensible to take the tes-
timony of the Masoretes more seriously when they claim to be
attempting to capture an established tradition of recitation with
their te‘amim, not creating one through the application of an es-
tablished grammar of the te‘amim.

As Dresher (1994, esp. 16-23) has explained, prosodic rec-
itations can be presented as a series of dichotomies, particularly
in regard to prose. But as Janis (1987, esp. 23-100) has also
shown, however, they need not necessarily be so understood.
Janis (1987, 48-53) has also shown that Wickes’s insistence that
nothing should break the “principle of continuous dichotomy”
can put the cart before the horse when it comes to understanding
the dynamics of prosody. More recent prosodic enquiry raises
new possibilities (Pitcher 2020). As Price (1990, 26-47) has
shown, the rules of the accents can explain almost all the ob-
served occurrences of te‘amim in the Twenty-One. This predicta-
bility extends to the poetic texts presented with the te‘amim of
the Twenty-One, but it must be admitted that more flexible zagef
te‘amim dominate these texts—that is, a ta‘am with more flexible
rules. The system of the Three was most probably created to ex-
tend this flexibility to short-verse poetic texts. In these texts, the
rules that appear to govern their distribution must either be un-
derstood to be very complex (so Wickes and Price) or to be held
more lightly—that is, in a position that is secondary to the poet-

ics. In both cases, it seems that the oral dynamics must be placed

Why Are There Two Systems of Tiberian Te‘amim? 319

to the fore if we are to make good sense of the te‘amim and that,
if so, two different sets of tecamim provide much interesting evi-
dence for the application of different ways of orally performing
different kinds of texts.

4.3. Ps. 18.19-20 as an Example

In Ps. 18 and elsewhere, it is possible to formulate a much sim-
pler and intuitive understanding of the te‘amim, by considering
how they impact the oral performance of the text. In the example
of vv. 19-20, for example, the anomalous lack of mid-stich dis-
junctive te‘amim in 19a, 19b, and 20a can be understood simply
to reflect a recitation tradition in which these stichs were recited
without any mid-stich “pause”. This absence thus causes the mid-
stich “pause” of 20b to be heard emphatically and so give its (se-

mantically) remarkable last clause a degree of special emphasis:
22 ww? MAT PR OPA ANTE
2 pam 3 a8)n! ap? vay

This gives an oral sense or, even, an oral taste to the text. In
transliteration, the effect can be seen when English punctuation

marks are used to represent the ‘pauses’ of the te‘amim:
yaqaddamiini bayom-’édi :_ wayhi-’ddondy lamis‘an Ii.
wayyosieni lammerhdb : yahallaséni—ki hdpes bi.

In English translation such a recitation might therefore be pre-

sented as follows:
They confronted me on the day of my trouble:

but the LoRD was there for my support.

320 Crowther

And He brought me out to the broad place:

He rescued me—because he delighted in me.

5.0. Conclusions

In an attempt to consider why there are two systems of Tiberian
te‘amim this paper has been compelled to explore a wide range of
observations. At this juncture it seems appropriate to draw them
together into a narrative that might explain why there are two
systems of Tiberian te‘amim and how they might relate to one
another.

Texts with Palestinian and Babylonian te‘amim employ the
same system of te‘amim for all the books of the Hebrew Bible. The
Tiberian use of two systems appears to be a Tiberian innovation
and not a phenomenon that was inherited by them. Early maso-
retic grammatical treatises consistently describe and praise the
Tiberian nequddot and te‘amim as recording, preserving and pass-
ing on an outstanding oral performance of the text. They do not
represent the te‘amim as a system of punctuation imposed upon
the text, but rather a way of presenting an outstanding oral reci-
tation of the text. A significant number of poetic texts in the He-
brew Bible are presented with the te‘amim of the Twenty-One,
most notably the books of Lamentations and Song of Songs. These
books evidence a modified use of the te‘amim of the Twenty-One.
The three books Psalms, Proverbs and Job stand apart as being
founded on parallelism and having significantly shorter verses.
The creation of a separate system of te‘amim was a response to

the combination of both the different oral dynamics of the reci-

Why Are There Two Systems of Tiberian Te‘amim? 321

tation of short verses and the different oral dynamics of the reci-
tation of poetry. The rules of the te‘amim of Three are more flex-
ible than those of the Twenty-One: simpler in general description,
but far more difficult to formulate in detail.** The mysteries of
the rules of the te‘amim of the Three only therefore appear to
resoluble when the focus is turned away from the rules of the
te‘amim and towards the dynamics of the recitation.

It has been found to be insightful to approach the te‘amim
of the Twenty-One as indicators of prosody as defined by linguis-
tics (suprasegmental phonology) in Pitcher (2020). The challenge
then remains before us to approach the te‘amim of the Three as
indicators of prosody as defined by poetics (which in this case
will be parallelism) and then, perhaps, to return to some of the
poetic texts of the Twenty-One—and their te‘amim—equipped

with new insight.

32 Price (1996, 1101) claims that for the te‘amim of the Three “Twelve
of the rules or auxiliaries operated without a single exception. The re-
maining rules operated with few exceptions and ranged in accuracy
from 94. 13% to 99.91%.” These impressive results, however, rely upon
his extensive use of ‘virtual’ te‘amim. As discussed above, the rules per-
taining to the transformations of ‘virtual’ te‘amim are not clear and the
rationale behind lists provided by Price is often very hard to discern.

322

Crowther

Appendix

Information relevant to Table 2, §3.5:

Gen. 49.2-27: Gen. 49.2, 5, 12, 14, 16, 19-21, and 23, [P3];
the rest, [P2] incl. v. 18 (three morphemes, one stich as per
BHS).

Exod. 15.1-18: Exod. 15.3, 5, and 14 [P3]; vv. 4 and 13,
four-stich lines delimited by ’atnah and tifha [P2]; vv. 1b-c
and 12, two-stich lines delimited by zagef [P1]; v. 18, two-
stichs delimited by tifha [P1]; the rest [P2].

Deut. 32.1-43: Deut. 32.1, 3, 5, 9, 12, 16, 18-19, 23, 26,
28, 29, 31, 33-34, 37, and 40 [P3]; the rest [P2].

Judg. 5.2-30: Judg. 5.18, 22, 25, 29 [P3]; v. 5, three (plus)
stichs delimited by ’atnah, zagef, and tifha [P2]; the rest
[P2].

1 Sam. 2.1-10: 1 Sam. 2.4, 6, and 7, [P3]; v. 2, three stichs
delimited by ’atnah and tifha, [P2]; the rest [P2].

2 Sam. 1.19-27: 2 Sam. 1.27, [P1]; the rest, [P2].

2 Sam. 22.2—50: 2 Sam. 22.8, 9, 31, 44, and 49, three stichs,
[P2]; vv. 3, 7, 16 four stichs, [P2]; the rest [P1]. V. 51 ex-
cluded, three stichs delimited by ’atnah and tevir?

Isa. 5.1-7: Isa. 5.3, four stichs delimited by ’atnah and tifha;
the rest [P2].

Isa. 40.1-31: Isa. 40.1, 13, 16-18, 23, 25, and 29-30 two
stichs delimited by ’atnah, [P3]; v. 8, four stichs delimited
by ’atnah and tifha, [P2]; the rest [P2].

Hab. 3.2-19: Hab. 2.5, 12, 15, and 18, [P1]; v. 7, three
stichs delimited by zagef and ’atnah, [P2]; the rest [P2].
Jon. 2.3-10: Jon. 2.9 [P3]; the rest [P2].

Why Are There Two Systems of Tiberian Te‘amim? 323

e Lam.: Lam. 1-2, [P2]; Lam. 3, [P1], Lam. 4, [P2]. Lam. 5,
[P1].

e Song: Song: the total stich count of P2 material is a matter
of some debate.

e 1 Chron. 16.8-36: 1 Chron. 16.29 and 33, three stichs de-
limited by ’atnah and zagef, [P2]; vv. 33 and 35-36, four
stichs delimited by atnah and zagef, [P2].

References

Baer, S., and H. L. Strack (eds.). 1970 (1879). Dikduke Ha-
Te‘amim des Ahron ben Moscheh ben Ascher. Jerusalem:
Magor (originally published Leipzig: Fernau).

Blau, Ludwig. 1896. ‘Massoretic Studies, Part 3: The Division into
Verses’. The Jewish Quarterly Review 9/1: 122-44.

. 1897. ‘Massoretic Studies, Part 4: The Division into
Verses (continued)’. The Jewish Quarterly Review 9/3: 471-
90.

Burrows, Millar, John C. Trever, and William Hugh Brownlee.
1950. The Dead Sea Scrolls of St. Mark’s Monastery. New
Haven, CT: American Schools of Oriental Research.

Cohen, Abraham. 1965. The Minor Tractates of the Talmud:
Massekhoth Ketannoth Translated into English with Notes,

Glossary and Indices. 2 vols. London: Soncino.

Cohen, Miles B., and David B. Freedman. 1974. ‘The Dual
Accentuation of the Ten Commandments’. In Masoretic
Studies 1, 7-19. Missoula, MT: International Organization

for Masoretic Studies Society of Biblical Literature.

324 Crowther

Crowther, Daniel J. 2015. ‘The Relevance of the Te‘amim to the
Textual Criticism, Delimitation and Interpretation of
Biblical Poetic Texts with Special Reference to the Song of
David at Psalm 18 and 2 Samuel 22’. PhD dissertation,
University of Bristol.

Davidson, A. B. 1861. Outlines of Hebrew Accentuation, Prose and
Poetical. London: Williams & Norgate.

Derenbourg, Joseph. 1871. Manuel du lecteur, d’un auteur inconnu
publié d’aprés un manuscrit venu du Yémen et accompagné de
notes. Paris: Imprimerie nationale.

Dotan, Aron (ed.). 1963. Sefer Diqduge ha-Te‘amim le-R’ Aharon
ben-Moshe ben-’Asher. Jerusalem: The Hebrew University of
Jerusalem.

(ed.). 2001. Biblia Hebraica Leningradensia. Peabody, MA:
Hendrickson Publishers.

Dresher, B. Elan. 1994. ‘The Prosodic Basis of the Tiberian
Hebrew System of Accents’. Language 70/1: 1-52.

Eldar, Ilan. 1992. ‘The Art of Correct Reading of the Bible’. In
Proceedings of the 9th Congress of the International
Organization for Masoretic Studies, 1989, 33-42. Atlanta:

Scholars Press.

Eldar, Ilan, and Yosef Ofer. 2018. The Masoretic Accentuation of
the Hebrew Bible According to the Medieval Treatise Horayat
ha-Qore. Jerusalem: Bialik Institute. [Hebrew]

Elliger, Karl, et al. (eds.). 1977. Biblia Hebraica Stuttgartensia.
Stuttgart: Deutsche Bibelstiftung.

Why Are There Two Systems of Tiberian Te‘amim? 325

Ginsburg, Christian D. (ed.). 1867. The Massoreth ha-Massoreth
of Elias Levita: Being an Exposition of the Massoretic Notes on
the Hebrew Bible. London: Longmans, Green, Reader & Dyer.

Heijmans, Shai. 2013. ‘Vocalization, Palestinian’. In Encyclopedia
of Hebrew Language and Linguistics, edited by Geoffrey Khan,
III:964-67. Leiden: Brill.

Jacobson, Joshua R. 2002. Chanting the Hebrew Bible: The Art of
Cantillation. Philadelphia: Jewish Publication Society.
Kahle, Paul Ernst. 1927. Masoreten des Westens: Texte und

Untersuchungen zur vormasoretischen Grammatik des

Hebrischen. 2 vols. Stuttgart: W. Kohlhammer.

. 1966. Masoreten des Ostens: Die altesten punktierten
Handschriften des Alten Testaments und der Targume.
Hildesheim: G. Olms.

Khan, Geoffrey. 2000. The Early Karaite Tradition of Hebrew
Grammatical Thought: Including a Critical Edition, Translation
and Analysis of the Diqduq of ’Abti Ya‘qtib Yusuf ibn Nuh on
the Hagiographa. Leiden: Brill.

. 2013. ‘Masoretic Treatises’. In Encyclopedia of Hebrew
Language and Linguistics, edited by Geoffrey Khan, II:598-
604. Leiden: Brill.

Janis, Norman. 1987. ‘A Grammar of the Biblical Accents’. PhD
dissertation, Harvard University.

Kittel, Rudolf, et al. (eds.). 1937. Biblia Hebraica. 3rd edition.
Stuttgart: Wiirttembergische Bibelanstalt.

Kugel, James L. 1981. The Idea of Biblical Poetry: Parallelism and
Its History. New Haven, CT: Yale University Press.

Levita, Elijah. 1538. Sefer Tov Ta‘am. Venice: Daniel Bomberg.

326 Crowther

Mercerus, Joannes. 1565. Sefer Ta‘ame ha-Migqra. Paris: G.
Morelius.

Pitcher, Sophia L. 2020. ‘A Prosodic Model for Tiberian Hebrew:
A Complexit Approach to the Features, Structures, and
Functions of the Masoretic Cantillation Accents’. PhD

dissertation, University of the Free State, Bloemfontein.

. 2021. ‘Towards a prosodic model for Tiberian Hebrew:
An intonation-based analysis. Stellenbosch Papers in
Linguistics Plus 63/1-27.

Price, James D. 1990. The Syntax of Masoretic Accents in the
Hebrew Bible. Lewiston, NY: Edwin Mellen Press.

. 1996. Concordance of the Hebrew Accents in the Hebrew

Bible. 5 vols. Lewiston, NY: Edwin Mellen Press.

Renz, Thomas. 2003. Colometry and accentuation in Hebrew
prophetic poetry. Waltrop: Hartmut Senner.

Revell, E. John. 1970. Hebrew Texts with Palestinian Vocalization.
Toronto: University of Toronto Press.

Sanders, James A. 1965. The Psalms Scroll of Qumran Cave 1
(11QPs‘*). Discoveries in the Judaean Desert 4. Oxford:
Clarendon Press.

Shoshany, Ronit. 2013. ‘Biblical Accents: Babylonian’. In
Encyclopedia of Hebrew Language and Linguistics, edited by
Geoffrey Khan, I:268-75. Leiden: Brill.

de Waard, Jan. 2004. General Introduction and Megilloth. Biblia
Hebraica Quinta. Stuttgart: Deutsche Bibelgesellschaft.

Wickes, William. 1881. A Treatise on the Accentuation of the Three
So-called Poetical Books of the Old Testament. Oxford: Clar-

endon Press.

Why Are There Two Systems of Tiberian Te‘amim? 327

. 1887. A Treatise on the Accentuation of the Twenty-one So-
called Prose Books of the Old Testament. Oxford: Clarendon
Press.

Yeivin, Israel, and E. J. Revell. 1979. Introduction to the Tiberian
Masorah. Missoula, MT: Scholars Press.

“SOME FANCIFUL MIDRASH
EXPLANATION”: DERASH ON THE
TE‘AMIM IN THE MIDDLE AGES AND
EARLY MODERN PERIOD

Benjamin Williams

Among the multitude of te‘amim ‘cantillation marks’ that adorn
the Masoretic Text of the Hebrew Bible, the accent shalshelet at-
tracts attention due to its conspicuous zig-zag shape and its sung
recitation as a trill or tremolo. Because of its rarity—it occurs just
seven times in the twenty-one prose books of the Hebrew Bible—
medieval and modern readers have attributed special significance
to the passages in which it appears. In his 1887 treatise on the
accentuation, William Wickes related medieval explanations to
the effect that the accent conveys information about the events
narrated not otherwise explicit in the biblical text, such as the
prolonged repetition of a particular action, or even angelic inter-
vention in the proceedings. Such aggadic interpretations were
not to the taste of sober-minded Wickes. Fearing that a similar
interpretation might underlie the Masoretes’ own use of
shalshelet, Wickes pronounced that the accent’s original meaning,
if it could be recovered, would not be worth the reader’s atten-

tion: “For we may be sure that we should have had some fanciful

330 Williams

Midrash explanation, which we can well afford to dispense with”
(Wickes 1887, 85).

The purpose of this study is to examine the history of the
idea that the shapes, names, and sounds of the te‘amim convey
information about biblical narratives. Medieval commentators
who relayed the peshat, the plain meaning of the text, regularly
employed the accents to identify pausal forms, stressed syllables,
the relationship between consecutive words, and the structure of
the verse. But a number of interpreters, including Tobias ben
Eliezer, Joseph ibn Caspi, Bahya ben Asher, and Moses Alsheikh,
also used them to formulate narrative details that are not explicit
in the text, including twists and turns in the plot, the thoughts
and motivations of the characters, and the manner in which di-
rect speech was delivered. The present study examines this tech-
nique first by analysing the midrashic method of deriving such
information from the graphic features of the consonantal text of
the Hebrew Bible. I will then turn to medieval anthologies of
midrash and commentaries that favour the derash, where unusual
and irregular cantillation marks, including shalshelet, are inter-
preted in a similar way. Finally, examples from the commentaries
of Moses Alsheikh of Safed (d. 1593) will show how sixteenth-
century Sephardi interpreters not only focused on exceptional
te‘amim, but treated the masoretic system of accentuation more
broadly as a source of information concerning biblical narratives.
As will be shown in the conclusion, medieval derash on the
te‘amim has inspired several contemporary expositors of the bib-
lical text. It is hoped that an impartial enquiry into the origins of

this exegetical method, which neither defends the interpretations

“Some Fanciful Midrash Explanation” 331

nor dismisses them as “fanciful,” will enable an understanding of
a distinctive interpretive approach to the Masora that has, once
again, become popular.

In Isaac Heinemann’s classic study of the midrashic
method, Darkhe ha-’Aggada, the significance accorded by the rab-
bis to the shapes and sounds of the consonantal text of the He-
brew Bible is designated as “creative philological” exegesis.
Though Heinemann focused on the interpretation of letters,
words, sentences, and sections, he acknowledged that other
graphic features of the text, including its division into para-
graphs, were also the subject of “philological” exposition (Heine-
mann 1970, 100). Interpretations of the puncta extraordinaria in
Sifre Numbers 69 illustrate this exegetical method. Among the
passages expounded is the reunion of Jacob and Esau in Gen.
33.4, where Esau fell upon his brother’s neck and kissed him. The

letters of 1nPw are written with supralinear dots:'

(1) span daQwH rN oy Ob nZANn ine p? Wy PTH
‘Esau ran to meet him. He embraced him, fell upon his
neck, and kisséd hirh, and they wept. (Gen. 33.4)

The midrash reads as follows:

(2) prea xd ix env ya pyaw '5 125 592 pwr xdw ,iApw 1a RNP
wad 522 PWH AYW AMINA pam Dai Xdx apy gw wyw

‘.,.An analogous case is “and kisséd hirh.” [The presence of
points above the word indicates] that [Esau] did not kiss
[Jacob] wholeheartedly. Rabbi Shim‘on ben Yohai said, “Is

1 Unless otherwise noted, biblical texts are cited from the BHS. The con-
sonants of the gere are printed in brackets.

332 Williams

it not certain that Esau hated Jacob? But at that particular
moment, his disposition changed and he kissed him whole-
heartedly.” (Sifre Numbers 69, ed. Kahana 2011-2015,
1:167)?

According to the first interpretation, the dots cast doubt on
the sincerity underlying Esau’s action. Shim‘on ben Yohai, by
contrast, suggests that the dots reinforce the significance of
Esau’s kiss as an indication of a profound change of heart. New
insights into the motivations and actions of biblical characters
may, according to these views, be disclosed by expounding the
text’s graphic features. This interpretation illustrates the relation-
ship Heinemann (1970, 13) held to be implicit between “creative
philology” and the resulting “creative historiographical” insights
into the narrative, since, according to the midrashic method, “the
interpretation of documents serves as a basis for the description
of history.’”*

Though the exposition of graphic features of the Hebrew
Bible’s consonantal text is well-attested in rabbinic literature
(Fishbane 2013, 17-21), a small number of references to maso-
retic signs can be found in late midrashim. An example comes in
the first part of Exodus Rabbah (2.6), which Avigdor Shinan

? Cf. Genesis Rabba 78.9. Midrash Tanhuma (printed) Va-yishlah 4 ex-
plains the insincerity of Esau’s action by suggesting that, rather than
seeking to kiss Jacob (from the root p"w3), he wished to bite him (from
J"wi). See also Liebermann 1962, 43-46; Shinan 1994; Martin-Contre-
ras 2003.

3 The full quotation reads: wi) :y7n2 739 Ara AT ow? bx m>yara lw
pon nsx pand w nmipa ppt by pp avon jiind oa waAwnN mitiynaz.

“Some Fanciful Midrash Explanation” 333

(1984, 23) has dated to the tenth century CE. The exposition of
Exod. 3.4, when God called Moses from out of the Burning Bush
by repeating his name, draws attention to other occasions when
patriarchs and prophets were similarly addressed. In the case of
Abraham, Jacob, and Samuel, the repeated proper nouns are di-
vided in the pointed Masoretic Text by a vertical bar (paseq).*
The midrash explains why the sign is not used in the case of Mo-

ses:

Ww APY? APY? ,Po|D 12. W OAIAN DATA :kvIA NN AWA AWN WRN
Pad pod 1a px awn awn Dar ,poa ia w Seinw Sxinw ,poa ia
ANWA pA Ap yyy wwnp1 dT NWA NNN Nn Xinw OTR
x> nwa oi ,oAny rato pan ox’ain 52 op sang nat hoya ar

py 52 proan

“And [the Lord] said, ‘Moses Moses’” (Exod. 3.4). You find
in the case of “Abraham, Abraham” (Gen. 22.11) that there
is a paseq. Likewise, there is a paseq in “Jacob, Jacob” (Gen.
46.2) and also in “Samuel, Samuel” (1 Sam. 3.10). But in
the case of “Moses Moses”, there is no paseq. Why is this
so? It is like a man who was laden with a heavy burden
and shouted, “So-and-so so-and-so, come over here and
take this load from me.”

Another interpretation (davar ’aher) is that God spoke in-
termittently with all [other] prophets, but never stopped
[speaking] with Moses throughout his whole life. (Exodus
Rabbah 2.6, ed. Shinan 1984, 116-17)

* See also Dotan (2005). An eleventh-century dating of this part of Exo-
dus Rabba has been advanced by Bregman (2003, 171-72). Cf. t. Be-
rakhot 1.14; Sifra Nedava parasha 1.12 (Weiss 3d); Genesis Rabba 56.7;
Tanhuma (Buber) Noah 1, 6, Va-yera 46, Shemot 15; Tanhuma (printed)
Va-yera 23, Shemot 18, Sav 13.

334 Williams

In good midrashic style, the darshan expounds Exod. 3.4 in
the light of verses throughout the biblical canon which exhibit a
similar syntactic formulation. Alternative explanations are pro-
posed, which, as indicated by the term davar ’aher ‘another inter-
pretation’, are not mutually exclusive (Fishbane 2013, 16, 21-
23). But, unusually for a midrash, the interpretation refers to the
masoretic pointing. The darshan’s observations correspond with
the text in the Leningrad Codex (dated 1008/9 cE), where a paseq
divides 07938 | 00738 ‘Abraham, Abraham’ in the account of the
Akedah (Gen. 22.11), ap” | 199° “Jacob, Jacob’ before the migra-
tion to Egypt (Gen. 46.2), and Dxinw | Sxinw ‘Samuel, Samuel’
when God called to the young prophet at Shiloh (1 Sam. 3.10).
The lack of a paseq when God called Moses’s name twice in Exod.
3.4, therefore, invites an explanation (Freedman 1998, fols 12a,
28b, 32b, 151b; Khan 2013, 10). According to the first interpre-
tation, the absence of the division that would indicate a slight
pause in the recitation means that God addressed Moses as hur-
riedly as someone shouting for urgent assistance with a heavy
load (Yeivin 1980, 216, no. 283). The alternative explanation re-
fers to the primacy of Mosaic prophecy, as Moses alone received
divine inspiration without interruption (cf. Leviticus Rabbah
1.14-15; Exodus Rabbah 21.4). By means of these explanations,
the darshan shows how the nature of the revelation at the Burning
Bush can be grasped through the midrashic interpretation of fea-
tures of the masoretic codex.

Expositions of the cantillation marks as sources of narrative
information can be found in the masora of tenth- and eleventh-

century manuscripts. The masora magna of the Aleppo Codex (ca.

“Some Fanciful Midrash Explanation” 335

930 CE) and of the Leningrad Codex compare the accounts of the
capture of two kings of Judah, Amaziah and Zedekiah, in 2 Kgs
14.13 and Jer. 34.21, respectively:

(3) Oster T2A wRin wan nts wRineTa Ta T2A in'yaN Hey

ww m3

‘And as for King Amaziah of Judah son of Jehoash, son of

Ahaziah, King Jehoash of Israel captured [him] at Beth-
Shemesh...’ (2 Kgs 14.13a)

(4) Oya YPIN TH PIR Ta HY PW NN TAATTIA PTE NI

QnA Yb 723 Tp 7TH

‘And as for King Zedekiah of Judah and his officials, I will

hand [them] over to their enemies and to those who seek

their lives, to the army of the king of Babylon, which has

retreated from you. (Jer. 34.21)

Though the first parts of the two verses are similarly
worded, the masoretic pointing differs. The initial nx in the ac-
count of Amaziah is pointed with the accent telisha. The -nxi in
the prophecy of judgement on Zedekiah, however, is joined by
maqgef to the following word and so lacks any accent and is
pointed with the short vowel segol rather than sere. The masoretic
note at 2 Kgs 14.13 in the Leningrad Codex explains the discrep-
ancy by relating Amaziah’s fate to the name of the accent telisha:

emaon> aim madann jo won pwrin .yon cmwm won pwrin
emadn> ain xdi miadan yn yoni pty

The former [i.e., Amaziah] [God] plucked (talash) and the
latter [i.e., Zedekiah] [God] snatched quickly. The former
was plucked (nitlash) from his kingship but returned to the

336 Williams

kingship. Zedekiah was snatched quickly from the king-

ship, but did not return to the kingship. (Freedman 1998,

fol. 211b)°

According to this interpretation, the masoretic pointing
communicates an element of the narrative. In 2 Kgs 14.13, the
telisha indicates that Amaziah was temporarily plucked (talash,
nitlash) from the throne. In Jeremiah, the short vowel on the
word ~nx} and its connection to 17*pTx ‘Zedekiah’ show that Zed-
ekiah’s downfall was quicker than Amaziah’s, since he was de-
ported to Babylon (2 Kgs 25.6-7) and never restored to the
throne (Loewinger 1960, 91-92; 1972, 603; Revell 2000, 72; Do-
tan 2009, 65-66; Ofer 2019, 261-63).

By the end of the tenth century, therefore, the midrashic
exposition of graphic features of the Hebrew Bible was no longer
limited to those of the consonantal text. Late midrashim interpret
masoretic signs, though not, to my knowledge, the names or
shapes of te‘amim. The masora itself derives narrative information
from the accents, though the verses discussed above are not ex-
pounded in extant midrashim (Friedeman 2021). But from the
late-eleventh century, certain midrashic anthologies and com-
mentaries developed insights into a small number of biblical nar-

ratives by explaining unusual te‘amim or anomalous patterns of

5 See also the masora magna of the Aleppo Codex at 2 Chron. 25.33, fol.
235b.

“Some Fanciful Midrash Explanation” 337

accentuation.° Several explain the account of Potiphar’s wife’s at-
tempt to seduce Joseph at Gen. 39.8, which begins with the rare

accent shalshelet:”

(5) WS 9D) MBIT AN YIN YTS 1 POTN WER TNH | 18071
oT 1g F7Wr

‘But he [Joseph] refused and said to his master’s wife,
“Look, my master has no concern, because of me, for house-

hold affairs, for he has entrusted everything he owns to

me.”’ (Gen. 39.8)

The earliest derash I have found on this ta‘am is in the late-
eleventh-century Leqah Tov of Tobias ben Eliezer, the Greek-
speaking exegete associated with the Byzantine city of Kastoria
(Ta-Shma 2005, 259-94; Mondschein 2009, 270-72; Cohen
2020, 166-67, 176-90). According to this explanation, the ta‘am
reveals the manner in which Joseph refused the advances of Pot-

iphar’s wife: ““But he refused.” Refusal upon refusal ad infinitum,

6 On the interpretation of further features of the Masora, see Penkower
(1982, xi, 31-40); Mondschein (2009, 270-72). On the interpretation
of tagin and irregular letters in the Sefer Torah, see Razhabi (1978, 90-
94, 120-23); Caspi (2015, 403-46). My thanks to Jen Taylor Friedman
for drawing my attention to Caspi’s study.

7 In addition to those discussed below, see also Gellis (1982-2014,
IV:94), and BnF MS Hébreu 5, fol. 1r. On the latter, see Wickes (1887,
85) and del Barco (2010, 42). On the interpretation of the Joseph nar-
rative in rabbinic texts, see Kugel (1990).

338 Williams

as it is written with pesiq and shalshelet...’ (Ben Eliezer 1884,
1:198)°

As Aron Dotan (1967, 164-65, 343-44) and Nurit Reich
(2006) have shown, shalshelet is also called marim, mar‘id, and
mesulsal in the Masora, names which characterise it as a distinc-
tive raising of the voice or as a trill or tremolo.’ Its association
with a loud or repetitive melodic motif would explain the com-
ment in the Leqah Tov. The shalshelet on jxi"1 ‘and he refused’
therefore indicates not only how the cantor should recite the
word, but also how direct speech was originally delivered and
that Joseph himself spoke with prolonged and insistent determi-
nation.'°

Several later exegetes used a similar method to explain the
verse. The fourteenth-century Provencal commentator Joseph ibn
Caspi (1280-ca. 1340), better known for his philosophical inter-
pretations of the Bible, included derash on the te‘amim in his
Masref la-Kesef (Mesch 1975; Twersky 1979; Herring 1982, 125-

8 spay tata .ndw>wai poaa nat ,ovAyD ADIN NN ANN PN .TNn1"
"DINDN PR MXN ITI ,puxnn. On the second part of the comment, ‘Re-
garding a sin, one must refuse; regarding a commandment, one must
not’, see Genesis Rabbah 87.5 and Yalqut Shim‘oni 145 (ed. Hyman,
1973, 750). On the paseq that always accompanies the shalshelet in the
twenty-one prose books, see Yeivin (1980, 188-89, no. 229).

° The shalshelet is also discussed in Goren (1989; 1995, 66-77, 151-56);
Morgenstern (1994).

© The comment is closely echoed in the Midrash Sekhel Tov (Ben Solo-
mon 1900-1901, I:239). On this work, see Cohen (2020, 193-205), the
afterword in Ta-Shma (2005, 253-94), and Mondschein (2009, 272-
77).

“Some Fanciful Midrash Explanation” 339

26; Ben-Zazon 2017, 87-95; Sackson 2017, 161-69).!! He wrote
that the shalshelet in Gen. 39.8 represents not determination, but
rather Joseph’s hesitation and wavering resolve in the face of

great temptation:

Awna n>w adran noi -wix-wnan gin ndbwdwn oy 03 1xn"1
Nya Uw Apa Wer 1A Dy 7D pRimANAN Sy nt onans 721
sing) ox 0795 Dx 7DIONA Paya Ata noi Dan ADP OX TAN pri
7a AAT ns TT np> ww www indir dizi pan 595 ww 7979
Fora my amt" pNdaw ati Sy mnxw 5"t [upmaqn ody.

.omnat noyn pond yprwon wer oxy x51 ayy wpa TAN

“And he refused.” The accent shalshelet is also among the
explanations that the Men of the Great Synagogue learnt
from Moses, about which I have already written regarding
the word mann} ‘and he hesitated’ (Gen. 19.16). [The ac-
cent] is not above the word jxn"1 on the two occasions it
refers to Jacob (Gen. 37.35; 48.19). There is no cause for
surprise if the wise man Joseph hesitated (noi) with re-
gard to this perilous matter, whether one way or the other,
for this befits every sage (and maybe his namesake took
another approach!).'* For how could anything be con-
cealed from our rabbis, of blessed memory, who said re-
garding Judah that an angel of the Lord was compelling
him, but regarding Joseph that he checked himself and
found that he could not [have intercourse]. Happy is the
one who can fully comprehend their sublime words!

1 On Ibn Caspi’s treatment of the te‘amim, see Rock (2007, §2.4). Iam
grateful to Dr Rock for kindly providing a copy of her dissertation.

” As suggested in the editions of Last (1905) and Rock (2007), this may
be a self-deprecating reference on the part of the commentator.

340 Williams

(Staats- und Universitatsbibliothek Hamburg MS Levy 8,

fol. 32b)"°
By crediting the transmission of the cantillation marks to the Men
of the Great Synagogue’ while also endowing them with Mosaic
authority, Ibn Caspi presents them as an authoritative source of
information regarding the biblical narrative. To understand the
significance of the shalshelet in question, Ibn Caspi refers the
reader back to his interpretation of Lot’s hesitant flight from
Sodom and Gomorrah (Gen. 19.16), where the word mannan) ‘and
he delayed’, is pointed with the same accent. There he explains
that the shalshelet’s meaning lies in its shape (ANW¥y2 Ary) and
that Lot’s indecisiveness was manifested physically as he “was
contorting his body (nnnyn vpyin nwiy) forwards and back-
wards.”!° The presence of the accent in Gen. 39.8 underlies Ibn
Caspi’s attribution of the same vacillation and tortuous hesitation
to Joseph, who, according to the interpretation in Midrash

Tanhuma and Genesis Rabbah, was saved from transgression only

‘3 This manuscript underlies the editions of Last (1905) and Rock
(2007), though the former prints a slightly different reading (ed. Last
1905, II:87-88).

™ Tbn Caspi frequently refers to the Men of the Great Synagogue when
explaining the accents, including in his comment on Gen. 1.1. The at-
tribution is in accordance with the rabbinic association of the events of
Neh. 8-9, including the reading of the Torah in such a way that it was
understood (Neh. 8.8), with the activities of the Men of the Great Syn-
agogue. See b. Nedarim 37b, b. Megillah 3a, and the texts examined in
Schiffer (1977). Cf. Bahya ben Asher’s assertion of the Mosaic origin of
the cantillation marks cited below.

1S MS Hamburg 8, fol. 23b; cf. Mishneh Kesef (ed. Last 1905), II:57.

“Some Fanciful Midrash Explanation” 341

through divine intervention, as the miraculous appearance of his
father’s image rendered him impotent.’© Ibn Caspi excuses Joseph
for his wavering resolve, recalling the principle that sages are
particularly susceptible to the evil inclination.’”

A third explanation is that of Ibn Caspi’s contemporary,
Bahya ben Asher of Saragossa.'® Bahya not only expounded the
shalshelet in Gen. 39.8, but also supplied an explanation for his
methods:

SITs Tad tax mat ndnn ote 77 PrTR nwE Ox KR TRAN

yean' nynaw oyom .> nwiy ns ant To awn ce xdm ada

spx7an mbans 1 pra xx yini inva Syn ata aor dp min

PIyya jal Andi now AN DYaa ux MInaw oNyoN TWA “nw
qa nna pti Dainnw oOTRAW Myunn

“But he refused and said to his master’s wife, ‘My master
is here [...].’” [Joseph] began by saying to her, ‘My master

16 This is related to the statement that “there was no man (vx 7X)”
present in the house with Joseph and Potiphar’s wife (Gen. 39.11) in
Tanhuma (printed) Va-yeshev 9 and Genesis Rabbah 87.7; cf. b. Sotah
36b and Rashi on Gen. 39.11. Cf. Levinson (1997, 279-81). Ibn Caspi
contrasts Joseph’s lack of resolve with that of his brother Judah, who,
according to Genesis Rabbah 85.8, approached Tamar only reluctantly
and through the coercion of the angel appointed over desire. Cf. the
interpretation in Solomon ibn Parhon’s Mahberet he-‘Arukh (1160-
1161) of the shalshelet on Gen. 19.16 as an indication of confusion
(1252). Ibn Parhon (fol. 5a); Berlin (1991, 85).

1” See b. Sukka 52a and also the ’aggadot of Rabbi Meir and Rabbi Akiva,
who were almost overcome by lust for the woman who turned out to be
Satan in disguise (b. Kiddushin 81a). Cf. Boyarin (2009, 258-66; Clen-
man (2014); and Rosen-Zvi (2011, 112-19).

8 On Bahya, see Walfish (1993, 216-17).

342 Williams

is here,’ which is to say, ‘Is my master not available to you?

What need do you have of me?’ And the cantillation mark

on the word jx70"1 shows that the matter was forbidden and

that he held himself back, refusing point blank. This is be-

cause we gain an understanding of what is not written in

the Torah from the cantillation marks, just as one may per-

ceive a person’s inner intention from his movements

(mun). (Ben Asher, ed. Chavel 1966, I:321)!?
According to Bahya, Joseph rejected Mrs Potiphar’s advances by
pointing to the immediacy of Potiphar’s presence with the words
hen ’adoni, “My master is here.” Potiphar’s availability to his wife
obviated any need of Joseph.”° His determination in refusing her
advances is indicated by the shalshelet. Bahya then details his
method of expounding the te‘amim as sources of supplementary
narrative information. His explanation hinges on a word play on
nyin, which refers both to ‘movement’ and ‘direction’ as well as
to the ‘vowels’ and ‘accents’ (Wolfson 1989-90, 1, 3; cf. Martini
2010, 61-65). Just as actions may speak louder than words, so
the accents that transform the biblical text into a dynamic me-
lodic motif disclose meanings that would not otherwise be appar-

ent.7!

Part of this comment was incorporated into the Minhat Shay, possibly
as an addition; see Norzi (2005-2006, 135).
20 Cf. Genesis Rabbah 87.5 and Tanhuma (printed) Va-yeshev 8.
"1 Bahya also justified his interpretation of the two te‘amim on tf ‘this
one’ in Gen. 5.29 as follows:
Do not think this matter is insignificant, since the whole
Torah is replete with allusions and matters of a philosoph-
ical nature (o»>2w ory). These were set forth providen-
tially in anticipation of the one who investigates the divine

“Some Fanciful Midrash Explanation” 343

Besides interpreting unusual te‘amim by means of derash,
our three commentators all refer to the accents’ conjunctive and
disjunctive functions and use them to determine stressed sylla-
bles.”* This is their principal significance in Rashi’s commen-
tary,”? where they are frequently used to identify stressed sylla-
bles, as well as the grammatical and syntactic functions of par-
ticular words.”* Abraham ibn Ezra likewise used the accents to

Torah. In this regard the sages explained that the cantilla-
tion marks in the Torah were also handed down from Sinai,
and they demonstrated this from what is written, “Giving
the sense so that they understood the reading” [Neh. 8.8].
They expounded this as follows: ‘“Giving the sense” refers
to the verses. “They understood the reading” refers to the
cantillation marks.’ (b. Nedarim 37b) (ed. Chavel 1966,
1:98).
On Bahya’s exegetical use of the method of sekhel, see Walfish (1993,
201-2); Talmage (1999, 319); Van der Heide (1983, 153).

22 See Leqah Tov on Exod. 13.11 (cf. Cohen 2020, 194-95); Ibn Caspi on
Gen. 1.1, 27; 3.23; 9.6; 18.21; and Bahya on Gen. 1.1; Exod. 25.38; Lev.
10.9 (on 1 Sam. 3.3); Lev. 23.16; Deut. 25.19; 32.5. As has been shown
by T. Cohen (1997-1998, 26, 43), even the accent shalshelet is accorded
no special significance in Ibn Caspi’s comment on Isa. 13.8, where he
follows David Kimhi in noting its disjunctive function (see the texts in
M. Cohen 1996, 98-99). I am grateful to Tamir Cohen for providing a
copy of his dissertation.

23 Existing studies include Englander (1939, 402-3; 1942-1943);
Shereshevsky (1972; 1982, 86-92); Kogut (1994, 42-54, 78-88, 148-
90); Himmelfarb (2004; 2005); Banon (2006).

4 Tt cannot be established with absolute certainty that Rashi did not
treat the te‘amim as sources of derash due to the lack of clarity regarding
the correct text of his commentary (Grossman 2012, 75-78; Lawee

344 Williams

parse words and explain syntax in accordance with his commit-
ment to grammatical exegesis,”* and there are numerous such in-
terpretations in the commentaries of David Kimhi.”° In contrast
to this common exegetical approach to the accents, derash on the
te‘amim is a relatively unfamiliar medieval method of exegesis,

being employed only in expositions that favour the derash and

2019, 15-20). However, the 45 comments on accentuation that I have
examined in Mikra’ot Gedolot ‘Haketer’, Bayerishe StaatsBibliothek Mu-
nich MS Cod. hebr. 5, and Fredman’s edition of the commentary on
Proverbs confirm that Rashi resorted to the accents to resolve questions
of grammar and syntax. Examples include Gen. 18.20; 29.6; 41.35;
42.21; 46.26; Num. 11.8; Deut. 11.30; Ezek. 40.18; Hos. 11.6; Ps. 10.3;
150.5; Job 18.20; Eccl. 3.16. The apparent lack of derash on the accen-
tuation could be explained by the absence of such interpretations in
Rashi’s sources of rabbinic exegesis. Cf. Kamin’s (1980, 24) argument
that, in Rashi’s biblical commentaries, “the root [w"77] in its various
forms indicates the source of the interpretation as taken from the
Sages”; see also Kamin (1986, 136-57); cf. Grossman (2017; 2021, 112-
14, 125-32, 256-81). Among the many studies of the relationship be-
tween midrashic interpretations and the plain meaning of Scripture
(xp Sw wow») in Rashi’s commentary, see Gelles (1981, 9-27, 42-65,
114-16); Ahrend (1997); Touitou (2000); Grossman (2017, 84-96); Co-
hen (2020, 95-126; 2021).

25 For instance, see his comments on Exod. 5.7; 18.3, 26; 29.35 (all in
the Long Commentary); Mic. 4.8; Nah. 1.1; Ps. 20.10; 45.6; 64.7. The
preface to Ibn Ezra’s commentary on the Torah includes criticism of the
methods of the Leqah Tov (1977a, I:7, 10); cf. Mondschein (2009, 271-
72). See also Wolfson (1988-1989, 3), and 86 of Ibn Ezra (1977b, 111).
Cf. Kogut (1994, 90-94, 196-230).

6 For instance, see his comments on Jdg. 6.16; 11.25; Isa. 28.17; 44.15;
Jer. 8.5; 9.18; 22.14; 22.20; 31.7, 36.20; Ezek. 15.4; 33.6; Ps. 35.19;
116.6. Cf. Kogut (1994, 56-57, 95-102, 231-38).

“Some Fanciful Midrash Explanation” 345

with reference to exceptional accents, such as the rare
shalshelet.?’

Bahya’s statement that one may “gain an understanding of
what is not written in the Torah from the cantillation marks” was
most likely known to Moses Alsheikh of Safed, who read and
cited Bahya’s commentary on Genesis,”* and who made full use
of this exegetical principle. Born around 1520, Alsheikh was of
the second generation of the Sephardi community that settled in
the Ottoman Empire (Alsheikh 1563, author’s introduction) fol-
lowing the expulsions from the Iberian Peninsula in the 1490s.
As shown in his many responsa, he was a student of Joseph Karo,
a communal rabbi, and a preacher. But Alsheikh is celebrated for
his biblical commentaries, an extensive corpus of exegesis that
covers almost the entire Hebrew Bible. His discursive, homiletic
style, and abundant use of midrash, have endeared him to gener-
ations of readers, and his commentaries remain popular to this
day (see Shalem 1965-1966).

7 Another example is the account of Lamech naming Noah, where two
accents appear on the word 7 ‘this one’ in Gen. 5.29 (Ben Eliezer 1884,
1:32, and Bahya’s commentary, as noted above, n. 25). The two accents
on 3139p ‘come near’ in Lev. 10.4 are expounded in interpretations at-
tributed to Judah the Pious and Eleazar of Worms; see the editions of
Konyevsky (1978-1981, 11:225) and Lange (1980, 42). On mystical in-
terpretations of the te‘amim, see Wolfson (1988-1989; 1989-1990); Dan
(1968, 70). On the interpretation of the accent shalshelet in the thir-
teenth- or early-fourteenth-century Sod ha-Shalshelet, see Idel (1988,
56-61); Fishbane (1994, 31).

8 See Alsheikh’s comments on Gen. 45.22; Prov. 30.29; Job 28.19; and
Song 5.8.

346 Williams

Like Tobias ben Eliezer, Joseph ibn Caspi, and Bahya ben
Asher, Alsheikh considered the meaning of the shalshelet in the
Joseph narrative. Ever the dutiful preacher, he formulated a mor-
alising interpretation that exhorts the reader to determined re-
fusal when faced with temptation, lest excuses or explanations be

undermined by the wiles of the tempter:

jor mim .nann byw nowhwo iaiwa nad (7x) pyran 72 by
iminad moa nws 18 pw 812 DWIX TIT 7D RINT WAIN Amann
xo TRap nwa AnanA ta myo pn a maT 72 NX> gm Map rat
nw> wr andan wa ap 72 ya nwyn wai nx dxnd Jad
Wk JRoinwia 725" opin Anan “aT mnw? mipon nawaw
row [a] p> ax wore Pn ox 72 72 AWY xd IpTpA Hd min
miuyo [7]> [0]3 770" ART O8 7D “ANNI NIP 7D DI OD DIwWl AWy"
WAY Wipw nd wo OoNTA ANWRI?D AoP pay WA AN An|NA TI
ND Jwn xdt D078 AIAN ...maye [ajax [qJ> [a]nsi nd>w5wa
nD oxw on qwad naw saa nend nxn adrian apan awyr px
Ta ware o[n]>bx> onxom ian ‘nd naw "AD DI WAR DWYER
ny ndwd>w imu ginw moa axa nbwbw opoa qran []nixa

xvas

“But he refused.” This means that he shook?” when he
heard it, like the shalshelet upon the word. Indeed, Joseph
behaved wisely. This is because it is human nature, when
an evil man or a foolish woman (cf. Prov. 9.13) comes to
entice [someone] to a sinful action that he does not intend
to do (cf. Isa. 10.7), that he will counter the tempter with

° On the meaning of n591 (Ruth 3.8), see Alsheikh’s comments on Deut.
3.29-4.1; Prov. 10.8; 12.17; Ruth 3.8; Job 6.18; and the introduction to
the commentary on Ecclesiastes. Given the definition in b. Sanhedrin
19b (cf. Targum Ruth 3.8) and the context of Joseph’s seduction, there
is also the possibility of double entendre. On humour in Alsheikh’s com-
mentaries, see his interpretations of Deut. 22.4—-5 and Ps. 49.2.

“Some Fanciful Midrash Explanation” 347

objections in order to silence him. But this is hardly the
way to save oneself from doing evil. For it might so happen
that the tempter is a smooth talker (cf. Ps. 140.12) who,
with flattering lips (cf. Ps. 12.3-4), will sway the fine
words of the one who is tempted, and he will be caught in
his net. But the one who has a brain in his head will not
act in this way, but rather will immediately resolve to re-
fuse and say that he will not do so under any circum-
stances, even if overpowered. Thereafter, if he so desires,
he can also list the objections to counter the tempter. This
is what happened in Joseph’s case. First, he made the re-
solve and “he refused,” like one who binds himself with a
chain (shalshelet). [Only] afterwards did he give the objec-
tions... This is the meaning of, “Look, my master [has no
concern, because of me, for household affairs, for he has
entrusted everything he owns to me. He is not greater in
this house than I am,] nor has he withheld [anything from
me except you, because you are his wife.] How could I do
this great evil [and sin against God (l-elohim)?]” (Gen.
39.8-9) being ungrateful to a human being, and thereby
also being ungrateful to the Lord. This is what is meant by
“and sin against God.” The same is indicated when it says
“and he refused” with shalshelet, to indicate that he puts
the chain of iniquity (shalshelet ‘avon) around his neck.
(Alsheikh 1593, fol. 65b)°°

Alsheikh begins by suggesting that the shape or melody of
the shalshelet indicates Joseph’s reaction to Mrs Potiphar’s ad-
vances—he trembled at the very thought. The ensuing explana-
tions hinge on the meaning of the word nbw5w ‘chain’. Alsheikh

associates Joseph’s exemplary decision to refuse temptation out-

3° The corrected reading jx" is from the 1710 edition, fol. 58a.

348 Williams

right with the accent, suggesting that he resolved to reject Poti-
phar’s wife as if bound by this ‘chain’ to his chosen course of
action. Alsheikh finally turns to Joseph’s commitment to proper
behaviour not only towards his master, but also towards God,
likening him to one who puts the my nb>w5w ‘chain of iniquity’
around his neck. This is the phrase that Rashi used to explain the
word 7p in the Babylonian Talmud (b. Shevu‘ot 31a), which
refers to the burden of personal responsibility that would be as-
sumed by a witness who testifies in a fraudulent case (see Berko-
witz 2006: 149, 278, n. 128; Sinai 2007). In suggesting that Jo-
seph’s words amount to a testimony, Alsheikh echoes midrashic
expositions of Gen. 39.9, “How could I do this great evil and sin
against God (l-elohim)?” as an oath by which Joseph committed
himself to shun the opportunity for sin.*' The shalshelet or ‘chain’
in the biblical text is the testimony to his vow before the divine
judge.

For all the creativity and ingenuity of his interpretations,
Alsheikh’s focus on the rare accent shalshelet as the key to under-
standing the narrative resembles the exegetical approach of the
medieval interpreters of Gen. 39.8 examined above. But Alsheikh
and other sixteenth-century Sephardi commentators of the Otto-
man Empire, including Abraham ben Asher and Solomon Alka-

bets, did not limit their expositions to a few exceptional

3! See the interpretation of Gen. 39.9 as an oath in Tanhuma (printed)
Va-yeshev 8; Genesis Rabbah 87.5; Leviticus Rabbah 23.11; Ruth Rab-
bah 6.4.

“Some Fanciful Midrash Explanation” 349

te‘amim.*” Rather, they saw the accentuation more broadly as a
source of information about biblical narratives. In order to exam-
ine this exegetical approach to the Masora, we will turn to three
comments in Alsheikh’s commentary on the book of Ruth, enti-
tled ‘Ene Moshe and first printed posthumously in Venice in 1601.
The commentary is structured as a series of discourses on ex-
tended pericopes. Each begins with a list of mbxw ‘questions’ or
nywip ‘difficulties’ which Alsheikh subsequently resolves. This
technique, for which Isaac Abravanel (1437-1508) is well
known, is ubiquitous in late-medieval and early-modern Sephardi
commentaries and homilies.*? A barrage of questions arouses the
reader’s curiosity about whether the text really makes sense and
whether the exegete can solve all the problems he has made for
himself. Alsheikh does so by examining the minutiae of the bib-
lical text, points he calls o:p17p7. His aim is to show that seem-
ingly trivial details, when properly understood, contribute to
overarching harmonious interpretations.

Alsheikh resorts to the te‘amim to solve exegetical problems

in the very first verses of Ruth:

(6) "wa? THT? ON? Tran WS TPA PISA ITM Dav vay "Ara HMI
"ALI IAW DY) TDN WNT DW P22 WWI INR] NIT ANIA Twa

32 On Abraham ben Asher’s interpretation of the zaqgef qatan in Gen.
12.1, presented in the course of his exposition of Midrash Genesis Rab-
bah 39.1, see Williams (2016, 75). On Solomon Alkabets, see his com-
ments on Ruth 1.11; 3.13, 17 (Alkabets 1992, 22, 188, 206).

33 See Bland (1990); Saperstein (2014a); Williams (2015); Lawee
(2008).

350 Williams

ANI TY INIT TT ON Ma DMN Fryo1 yma | MATA’ Oy

OW PT
‘And it came about (771) in the days when the judges judged
that there was (71) a famine in the land. So a man of Beth-
lehem of Judah (97377 on) nan) went to reside in the fields
of Moab, he and his wife and his two sons. The man’s name
was Elimelech, his wife was Naomi, and his two sons were
Mahlon and Chilion. They were Ephrathites from Bethle-
hem of Judah (AT7 on? man). They came to the fields of
Moab and were there.’ (Ruth 1.1-2)

Alsheikh begins by enumerating no fewer than ten nrwip ‘diffi-
culties’ regarding these verses, asking why *7" ‘and it came about,
there was’ and nim on? na ‘Bethlehem, Judah’ are repeated, and
why the family members are introduced once anonymously and
then again by name. The eighth difficulty focuses on how
Elimelech is introduced in verse two:
TAI [AJA TM wT OW TAIN 'n...kopAa pand wana van
2 TDR

The following must be understood in this passage of Scrip-
ture... 8. The statement ‘the man’s name [was Elimelech],’
as it would have sufficed to say ‘his name was Elimelech.’
(Alsheikh 1601, fol. 3a)
Alsheikh here calls attention to an apparent tautology. Revealing
his conception of Scripture as marked by perfect felicity of ex-
pression, in which no detail is superfluous, he asks why Ruth 1.2
states 72°>R wero ow when 72x inw would have been more

concise.

“Some Fanciful Midrash Explanation” 351

Alsheikh’s explanation revolves around two concerns:
Elimelech’s social status and the halakhic question of the circum-
stances in which one is permitted to leave the land of Israel. This
latter is discussed with reference to the book of Ruth in b. Bava
Batra 91a. On the one hand, Elimelech’s departure at a time of
famine suggests that scarcity of food is a permitted reason to
leave the land of Israel. On the other hand, he and his sons die in
the next three verses, suggesting that departure even in the direst
of circumstances is forbidden.** Alsheikh seeks an explanation
partly in the talmudic principle that “the Holy One, blessed be
He, is exacting with his righteous ones to the extent of a hair’s
breadth” (Cf. b. Yevamot 121b; y. Sheqalim 48d (5.1), y. Betsa
62b (3.8); b. Bava Qamma 50a). Thus, even if departure from the
land of Israel is tolerated in particular circumstances, Elimelech’s
social status meant that he was held to particularly high stand-
ards. But to demonstrate this, Alsheikh must show that Elimelech
was indeed important or righteous, a detail not explicit in Scrip-
ture. He alludes to the rabbinic interpretations that Elimelech
and his sons were “great men of their generation” and “leaders
of their generation” (b. Bava Batra 91a; Ruth Rabbah 1.4) and
adds insights of his own:

wri aad. wern owi72 mn di ot xdn°2 pana Sy annn ox

x"na on Sopnaw [opwix 599 5173 RINW WR ANIN NInW DWI

mada sans aoe maw >"? [ax]w [iad Tox oa ayy
nat iniaa bxaw miadnd wn prw ax mn inindw an dpw

34 See also Sifra be-Har parasha 5.4 and Moses Maimonides, Mishneh
Torah, Hilkhot Melakhim u-Milhamotehem 5.9-12. Cf. Encyclopedia Tal-
mudica, s.v. ‘Erez Israel’, III:47; Safrai (2018, 78-79); Kanarfogel
(1986); Saperstein (2014b, 281).

352 Williams

Ty ony pIptn n"apn yao? Spw ...5"173.4"1b oyy xin InaAN
121 Doms nian ni yawn Wwawi ...w 773

Do not be surprised at the matter, for was [Elimelech] not
a great man? This is because ‘the name of the man [was
Elimelech]’ (72°98 wx DWI) means ‘the designated man
[was Elimelech]’. This is a way of describing a man as
‘great’, like all [who are styled] 0’w3x in Scripture. The use
of the definite article also [indicates this], as does [the
name] Elimelech. [This is] as the sages said, ‘[Elimelech]
would say, “Kingship will come to me (m125n ...75x).”” Be-
cause of his pre-eminence he would say that no one was
better suited for the monarchy of Israel than he was. And
to indicate his exalted position is the cantillation mark
pazer gadol... On account of this, the Holy One, blessed be
He, was strict with them to the extent that... they were
sentenced to death. (Alsheikh 1601, fol. 4a)

Alsheikh demonstrates that each word of the phrase wxn ow
q2mx indicates Elimelech’s high standing. ow shows that he is
singled out as an important individual. Midrashic interpretations
of the word wx treat individuals so designated as particularly
righteous, such as the exposition in Genesis Rabbah 30.7 of Noah,
the “man righteous and wholehearted” (Gen. 6.9).*° Elimelech’s
name itself indicates his aspirations. Alsheikh relates the inter-

pretation in Ruth Rabbah 2.4 that revocalises his name to show

3 “Wherever the word ’ish occurs, it refers to a righteous man who fore-
warned [his generation]” (Theodor and Albeck eds. 1903-1936, 272).
Cf. Numbers Rabbah 16.5. For Alsheikh, the same applies to nWx, and
he interprets the designation of Rebekah as nWx7 with the definite arti-
cle in Gen. 24.39 (ad. loc.) as an indication of her importance.

“Some Fanciful Midrash Explanation” 353

that he positioned himself to become Israel’s first king by claim-
ing that “kingship is mine”, min ‘x.

Alsheikh supports these interpretations by referring to a
feature of the biblical text itself: the ta‘am on Elimelech’s name.
Though the disjunctive accent pazer is not unusual (it occurs 858
times in the prose books of the Hebrew Bible), it appears only
here in Ruth (Price 1996, I:5, IV:831). In the public recitation of
the book, the melodic motif unique to this verse and the pause
indicated by the accent draw attention to Elimelech’s name at the
moment he arrives on scene. Alsheikh refers to this accent as
pazer gadol, a name which holds the key to the interpretation that
it “indicates [Elimelech’s] exalted position”: a ‘great pazer’ her-
alds the entrance of the great Elimelech.*° It thus helps to explain
the significance of the expression 77/°9x wx DW) and supports
the overarching interpretation that, due to his importance, he
was held to high standards and punished for leaving the land of
Israel even at a time of famine.

Alsheikh resorts to the te‘amim again in his comment on the

narrative of Ruth gleaning in the field in chapter 2:

(7) Ayan Tyan wh cnt TIyaIn +? OIpAT oY ayaa Aya? 1a TaN
‘agin TTWA APIO NIWA] NT MAND TWIT OYA oy

3° In his commentary on Lev. 23.27, Alsheikh similarly designates the
pazer on the word 78 as pazer gadol; he does not use the term to refer to
qarne farah (see Yeivin, 1980, 212-13, nos. 274-76). The interpretation
of Ruth 1.2 is analogous to that of Est. 6.7, where the zagef gadol on the
word wx indicates the great importance of the individual concerned.

354 Williams

‘Boaz said to his servant who was stationed over the reap-
ers, “To whom does this young woman belong?” The serv-
ant stationed over the reapers answered and said (...}D*"1
7x1), “She is a young Moabite woman, the one who re-
turned with Naomi from the fields of Moab.”’ (Ruth 2.5-6)

Alsheikh begins with the characteristic litany of questions.
Among them, he asks why two verbs introduce the servant’s re-
ply, v4 and then 79X", when one would suffice:

SD agin MyI Tax ad A nA ANN yd os ANI...

...In addition, the word ‘and he answered?’ ({p*1) appears to

be superfluous as it would have sufficed to say, ‘And he

said (Jax), “She is a young Moabite woman.” (Alsheikh

1601, fol. 17a)
This question prompts an elaboration of the narrative. Alsheikh
explains the role of the servant, his relationship to Boaz, and the
particulars of their exchange. Because the servant was appointed
or stationed “over” the reapers, Alsheikh describes him as stand-
ing on a platform to survey the harvest. He also develops inter-
pretations from the Babylonian Talmud (b. Shabbat 113b) and
Rashi’s commentary (on Ruth 2.5), that Boaz asked about Ruth
not because he habitually enquired whether young women were
single, but because he noted how carefully she observed the ha-
lakhic regulations about gleaning. In the hands of Moses
Alsheikh, this rabbinic interpretation germinates into an ex-
tended narrative in which the servant misinterpreted Boaz’s in-

tentions and so embarked upon a character assassination of Ruth

“Some Fanciful Midrash Explanation” 355

to prevent his master from becoming entangled with a Moabite

woman:?”

Sa poawipn bp aya wer ayad mitw dpa pot man ‘an apin wn
wer vaw 525 nd nx ART mA DIpAl Trnyrd) onaXdnl iM
Sip ond 12 na Wwe wr pays bon ims inadi xy x5 wwe ofA
Do IN Wwiys owyi dT Wiad AAA NTWwa or nA DpINdi Danpd
mina 1905 Sip od wad d qaKR1 yp Ip IN’ WR DID
781 APR PA MAKI Mayr ina yp naan pwdn apan ps aan nn
yann gin adyndn poy wwe oyon xin ad tyo nypi ta) oP Tar?

pian

“And the servant answered and said...” The practice of
field owners is to station a man appointed over the reapers
so that they do not get lazy in their work. [They] station
him in an elevated place [from which] he can see them all,
so that no one will let [any grain] drop without him seeing
it. [Owners] choose this individual from among all their
servants, someone who has the strength to raise their voice
to those near and far. This is particularly [important] in
the fields of a great man like Boaz who had immense
wealth and property, for many [people] were reaping his
harvest with him. And it says that, in order to reply to
[Boaz], he raised his voice to denigrate Ruth. This is the
meaning of ‘and the servant answered [and said]’ (....p%
7px). The expression indicates that he raised his voice,
just as in the case of, “And you will answer and say ( min
manx}) [before the Lord your God, ‘An Aramaean was seek-
ing to destroy my father...’]” (Deut. 26.5) and, “And Job
answered and said (7781 1x yw), [Let the day on which
I was born perish...’]” (Job 3.2). And a little support for

3” Contrast with the overseer’s words in Ruth Zuta 2.7 and Targum Ruth
2.6, where he points out that Ruth is a convert.

356 Williams

this may be [drawn from] the cantillation mark on the pe-

nultimate syllable, which is revia‘. This is the meaning of

yw. (Alsheikh 1601, fols 17b-18a)
According to Alsheikh, the two verbs jp and 7X") indicate that
the overseer spoke loudly. Deut. 26.5 and Job 3.2 both introduce
direct speech in this way, and Rashi’s commentary explains on
each occasion that the phrase indicates a raising of the voice.*®
Alsheikh appeals to the ta‘am on jp*1 to show that this interpreta-
tion holds true in the verse in question. This is one of several
occasions in his commentary where he focuses on the melodic
function of the disjunctive accent revia‘. Elsewhere he describes
it as IMKN1 dip may ‘like one who raises the voice’ to communi-
cate a particular interpretation.*’ Here it appears on the first
word of the verse and introduces direct speech. Alsheikh there-
fore suggests that the accent indicates how the ensuing statement
was delivered and that the servant shouted out an urgent warning
to Boaz. This interpretation is in accordance with the exegetical
technique observed above in the Leqah Tov, which treats the can-
tillation marks both as musical signs for the cantor and as indi-
cations of how direct speech was originally delivered by biblical
characters. By supporting the interpretation that the overseer was
shouting, the revia‘ helps Alsheikh to formulate a narrative that

answers his initial question about an apparent tautology. He

38 See the 1546-1548 Rabbinic Bible (Venice: Bomberg), fols 216b,
785a, and the texts discussed in Smelik (2013, 58-67).

39 See, for instance, Alsheikh’s comment on Gen. 24.7. Cf. Rashi’s com-
mentary on Gen. 1.1 and 37.20, where significant phrases ‘speak’ to the
expositor, saying 71w77 ‘expound me’ (Ben Isaac, 1982, 2, 134); regard-
ing the latter, cf. Tanhuma (Buber) Va-yeshev 13.

“Some Fanciful Midrash Explanation” 357

shows that there is no redundancy in the use of both qv and 77x",
as the accent on the former reveals its distinctive shade of mean-
ing.*°

A final comment on the te‘amim concerns the exchange be-
tween Ruth and Naomi after the harvest. The third chapter of

Ruth begins:

(8) Amy 77a Ws Ni Tz WPAK NVA Ha ANN *pyI A? WDNM
DywA Mans ADP NIaThan Powis mea Wwe anya twa xa
THA” (K) Tn[7aw] (Q) naw (K) ‘naw NpeI | MYM ANA

ni?)
‘Naomi her mother-in-law said to her, ‘My daughter (*na),
should I not seek security for you, that you may be well?
Now is not Boaz, with whose young women you were, our
kinsman? He is about to winnow the barley at the thresh-
ing-floor tonight. Now wash, anoint yourself, put on your
cloak and go down to the threshing-floor. Do not make
yourself known to the man until he has finished eating and
drinking.’ (Ruth 3.1-3)

Among the nrwip, Alsheikh lists the following:

“© A variant pointing of the word with the accent darga is attested in a
number of manuscripts; see Wright (1864, 9 [second pagination]). How-
ever, the explicit reason for Alsheikh’s reference to the accent of qv‘: is
to support his account of the overseer’s actions, rather than to assert the
correct reading of the text.

“ As printed in the 1546-1548 Rabbinic Bible (Venice: Bomberg), fol.
831a. See the footnote below regarding the pointing of *nha.

358 Williams

Seinw 125 nw xd pao proaona aNipw w> and pea nn on...
pan ona 20> pr wwe oat ean > ainad wap nna x20

...Furthermore, what is the use of telling us that [Naomi]

called [Ruth] “my daughter” (na)? For without a doubt

the prophet Samuel did not intend, by means of the holy

spirit, to write lots of words for us which serve no purpose

for the narrative. (Alsheikh 1601, fols 28a—28b)

Alsheikh here makes explicit a key assumption underlying
his interpretations. Referring to the talmudic attribution of the
book of Ruth to Samuel (b. Bava Batra 14b), he accords it the
status of an inspired prophetic writing. This means that nothing
is redundant and, as he asserts, every textual detail contributes
to the book’s narrative. In the comment that follows, this princi-
ple is applied both to the word *na and to its accent.

Alsheikh refers to a kabbalistic interpretation related by
Nahmanides and the Midrash ha-Ne‘elam on Ruth. When
Nahmanides expounded Onan’s failure to raise up offspring for
his late brother, he referred to levirate marriage as “one of the
great secrets of the Torah.” Concealing the nature of this “secret”
from the casual reader, Nahmanides referred allusively to Ruth
4.17 and stated pa S2>wnm ‘and the wise will understand’ (com-
mentary on Gen. 38.8 in Ben Nahman 1959, I:214-15; see Idel
1983; Wolfson 1989; 1993; Yisraeli 2006). This is a reference to
the women of Bethlehem, who celebrated the birth of Obed by
saying not “a son is born to Ruth” but rather “a son is born to
Naomi.” As explained in the Midrash ha-Ne‘elam, this indicates
that Ruth’s son was in fact the reincarnation of her late husband
Mahlon (Midrash ha-Ne‘elam on Ruth, ed. Margaliot, 2007-2008,

“Some Fanciful Midrash Explanation” 359

89d-90a).” The “great secret” of levirate marriage, it would
seem, is that the soul of the deceased is reborn in the child be-
gotten of the union.

Because Ruth’s marriage to Boaz would secure Mahlon’s re-
incarnation, Naomi had a vested interest. This calls into question
her motivation in arranging the rendezvous at the threshing
floor. Alsheikh defends Naomi’s altruism by explaining the word
*Aa and its accent:

anda *ayi qaxm [Jax nasa by wpnynd wipn nn axa 7 by
AIT pRw ANN ANY op nAwl ANwWy Mwynsa nay XT AWK
md min ar ya anion ays n> aaxni inn anda ndyin ans aim
nbna Syma apr oyo nvn pp on arena ox aK. wn oyia
wN WRI Prysa WAX ox qa xin1 AT NINA Moy nd ton ona
qnida *na ox 7D oxya nx nda xdoD And> oY niand JNK myn

Jngin wpar xb and awnn > naa Janpaa nn

Therefore, the holy spirit came to show us the truth of the
matter and said, “Naomi said”. This indicates that she was
as pleasant (7n’y3) in her actions as her name suggests even
though, being [Ruth’s] mother-in-law, it was not in her na-
ture to pursue the benefit of her daughter-in-law. That is
the meaning of, “Naomi her mother-in-law said to her.”
[Naomi] demonstrated this to [Ruth] with her pleasant
words (7572 nya) when she said to her, “My daughter
(ona)...” The cantillation mark zagef gadol on the word "na
must also be examined precisely (p1p17’) because it [like-
wise] indicates that Naomi had such [an attitude] towards
Ruth. It is as if to say, “When I instruct you, do not think

” Cf. Zohar Mishpatim 2.99b and Zohar Va-yeshev 1.188a-—b (ed. Matt,
Wolski, and Hecker 2004-2017, III:148-50; V:38; XI:263-65). See
Mopsik (1987, 16-21); Hallamish (1999); Fine (2003, 304-14); Wer-
blowsky (1997, 112-15, 234-56).

360 Williams

of me as a mother-in-law [talking] to her daughter-in-law,

for you are not in fact my daughter-in-law, but my daugh-

ter. This is to say that the spirit of my son is inside you,

and you should be considered as a daughter to me. Why

would I not seek your benefit?” (Alsheikh 1601, fol. 28b)

This comment is an atomistic reading of the words ‘nyi
‘na mAninn. Naomi’s name is mentioned explicitly in order to
evoke the etymology in Ruth Rabbah that she was “pleasant
(nay1) in her actions” (Ruth Rabbah 2.5; 3.6). Her kind-hearted-
ness prevailed over what Alsheikh considers to be the nature of
the mother-in-law, who does not pursue her daughter-in-law’s
best interests.*? This insight is supported by the word *na and its
accent. In interpretations similar to that of the aforementioned
revia‘, Alsheikh likens the zagef gadol in his commentaries on
Eccl. 9.10 and Est. 6.7 to “one who raises the voice” to declare a
particular interpretation. In this case, the disjunctive accent on
the initial word of Naomi’s speech focuses attention on the ex-
pression that captures the true relationship between the women.
Alsheikh rewrites Naomi’s words to show that she considers Ruth
her daughter and treats her accordingly. The word *na and its
accent thus support the interpretation that Naomi arranged
Ruth’s liaison with Boaz purely out of concern for Ruth’s wellbe-

ing rather than as a selfish means to secure Mahlon’s rebirth.*

43 On the portrayal of the mother-in-law in rabbinic texts, see Ilan (2017,
120-22).

“4 A variant pointing of *na with the accent revia‘ is attested in a number
of manuscripts and printed editions, though, as noted in the case of jp*1
above, Alsheikh draws no attention to the different reading. See Gins-

“Some Fanciful Midrash Explanation” 361

By means of these three comments on the accentuation of
Ruth, Alsheikh has opened up remarkable new vistas on a well-
known narrative. Elimelech’s pazer reveals his high social status.
A revia‘ shows that Boaz’s servant was stationed on a platform,
shouting to the harvesters and, in an unfortunate misunderstand-
ing, even to his master. And a zagef gadol shows that Naomi
shunned selfishness and spoke to Ruth out of maternal compas-
sion. These te‘amim disclose information about the narrative that
is not otherwise indicated in the text. In this respect, Alsheikh’s
comments resemble the aforementioned interpretations of Gen.
39.8, in which the Legah Tov, Ibn Caspi, and Bahya ben Asher
derived the details of Joseph’s refusal of Potiphar’s wife from the
accent shalshelet.

An important difference between Alsheikh and his prede-
cessors is that, rather than explaining peculiarities as he encoun-
ters them, he goes in search of te‘amim that might serve as useful
sources of narrative information. The accents he selects in Ruth
are not unusual in themselves, though in a particular verse, pazer,
revia‘, and zagef gadol stand out from the most familiar sequences

of te‘amim.*° Alsheikh ascertains their meanings from their names

burg (1926, 579); Wright (1864, 16 [second pagination]). In the Lenin-
grad Codex (fol. 422a), the word is pointed with gershayim. This is also
the reading of the 1601 edition of Alsheikh (1601, fol. 27b), in which
the biblical text printed alongside the commentary obscures the mean-
ing of Alsheikh’s interpretation. On the significance of the accents and
Zoharic references to the masoretic pointing among early modern Kab-
balists, see Penkower (2010); Dweck (2011, 151-69); Rubin (2011).

45 Revia‘ occurs 8910 times in the prose books; zagef gadol 1655 times
(Price 1996, I:5).

362 Williams

and their melodic functions. But his discussions also draw on in-
sights from midrashim, Rashi’s and Nahmanides’s commentaries,
and the Midrash ha-Ne‘elam. Alsheikh uses the accents to support
these explanations by showing that they may be derived directly
from features of the biblical text. One reason for this is evident
in the comment on the word *mAa, where Alsheikh states that the
zagef gadol should be “examined precisely (pt oyv nvn ptpiTy
511)”. As mentioned earlier, forms of p1775 are used by Alsheikh
and other contemporary Sephardi commentators to refer to the
scrutiny of the biblical text to find answers to the nrwip, the ques-
tions raised in the pericope. That Alsheikh used the te‘amim to
this end was already apparent in his comments on the pazer on
Elimelech’s name and the revia‘ on }p*1. Both respond to questions
about seemingly superfluous words. But the interpretation of *na
makes explicit that Alsheikh counts the accents among the minu-
tiae of the biblical text which, properly understood, demonstrate
its overall coherence.

By appealing so readily to the te‘amim, Alsheikh treats the
accents as an essential and fundamental means by which biblical
narrative is expressed. No longer are they a paratextual guide to
the grammar and syntax of the words; nor are they occasional
indicators of unexpected interpretations. Now they are treated as
an integral part of the text itself, conveying information that is
necessary to understand the narrative with clarity. The reader of
the biblical text must therefore be constantly alert to the bearing
that every accent, however commonplace, might have on the
course of events in any given passage. This manner of reading the
Hebrew Bible was enabled in many editions of Alsheikh’s works

“Some Fanciful Midrash Explanation” 363

that were issued in Venice by Giovanni di Gara, including the
1601 editio princeps of the commentary on Ruth, by the provision
of a vocalised and accented text alongside the commentary.”°
This mise en page allows the reader to move from an encounter
with the accented words of the biblical text to Alsheikh’s ques-
tions regarding their significance and coherence, and finally to a
problem-solving exegetical discourse that shows how studying
the details of the accented text allows one to grasp its full mean-
ing.

The idea that the te‘amim indicate not only grammar and
syntax, but also narrative information has resurfaced in several
recent expositions of the Hebrew Bible. In their homilies on the
Joseph narrative, Louis Jacobs, Jonathan Sacks, and Jonathan
Magonet find common cause in interpreting the shape and quiv-
ering tone of the shalshelet in Gen. 39.8 as an indication of the
protagonist’s inner conflict, struggle, torment, and crise de con-
science. For Jacobs (2004, 59-60), the ta‘am “expresses vacilla-
tion where we would expect firm resolve” and, for Magonet
(2004, 27-28), Joseph was “fighting against the temptation to
accept.” Sacks (2009, 109-15) ascribes his interpretation to an

“© Partly through the efforts of Alsheikh’s son Hayyim, Di Gara issued
Alsheikh’s commentaries on Daniel (1591), Song of Songs (1591; 1606),
Proverbs (1601), Ruth (1601), Lamentations (1601), Qohelet (1601),
Job (1603), and Psalms (1605) with the biblical text printed alongside.
This typographical complication was omitted from the earlier editions
of the commentaries on Song of Songs (1563) and Daniel (1563) that
were printed in the Ottoman Empire. The list of printed editions of
Alsheikh’s commentaries compiled by Naphtali Ben-Menahem is in
Shalem (1965-1966, 237-74). See Benayahu (2001); Dweck (2010).

364 Williams

earlier exegete, Joseph ibn Caspi, developing his predecessor’s
interpretation that the ta‘am indicates a physical manifestation
of wavering resolve by suggesting that it reveals “a psychological
state of uncertainty and indecision.” As the cantillation marks
once again “raise their voices” to relay interpretations old and
new, it is hoped that an understanding of the development of this
expository technique and its relationship to earlier exegetical
methods will enable a deeper appreciation of a chapter of the
reception history of the Hebrew Bible, in which the Masora is
treated as a means to “gain an understanding of what is not writ-
ten in the Torah” (Ben Asher, 1966, I:321).

References

Ahrend, Moshe. 1997. ‘L’adaptation des commentaires du Mid-
rash par Rashi et ses disciples a leur exégése biblique’. Re-
vue des Etudes Juives 156: 275-88.

Alkabets, Solomon. 1992. Sefer Shoresh Yishai: Perush ‘al Megilat
Rut. Jerusalem: Hivzon.

Alsheikh, Moses. 1563. Havaselet ha-Sharon. Constantinople:
Jabez.

—. 1593. Sefer Torat Moshe. Belvedere: Ashkeloni.

—. 1601. ‘Ene Moshe. Venice: Di Gara.

. 1710. Sefer Torat Moshe. Amsterdam: Proops.

Banon, David. 2006. ‘L’exégése de Rachi sur les té’amim’. In Hé-
ritages de Rachi, edited by René-Samuel Sirat, 127-36.
Paris: Eclat.

del Barco, Javier. 2010. Bibliothéque nationale de France: Hébreu 1

a 32: Manuscrits de la bible hébratque. Turnhout: Brepols.

“Some Fanciful Midrash Explanation” 365

Ben Asher, Bahya. 1966. Be’ur ‘al ha-Torah. Edited by Charles
Chavel. 3 vols. Jerusalem: Mossad Harav Kook.

Benayahu, Meir. 2001. ‘Rabbi Isaac Gershon’. Asufot 13: 9-89.
[Hebrew]

Ben Eliezer, Tobias. 1884. Lekach-Tob. Edited by Solomon Buber.
2 vols. Vilna: Romm.

Ben Isaac, Solomon. 1982. Perushe Rash"i ‘al ha-Torah. Edited by
Charles Chavel. Jerusalem: Mossad Harav Kook.

Ben Nahman, Moshe. 1959. Perushe ha-Torah. Edited by Charles
Chavel. 2 vols. Jerusalem: Mossad Harav Kook.

Ben Solomon, Menahem. 1900-1901. Midrash Sekhel Tov. Edited
by Salomon Buber. 2 vols. Berlin: Ittskovski.

Ben-Zazon, David. 2017. ‘““Parashat ha-Kesef” le-R. Yosef ibn
Kaspi’. In Five Early Commentators on R. Abraham Ibn Ezra,
edited by Howard Kreisel, 87-95. Beer-Sheva: Ben-Gurion
University of the Negev Press.

Berkowitz, Beth. 2006. Execution and Invention: Death Penalty Dis-
course in Early Rabbinic and Christian Cultures. New York:
Oxford University Press.

Berlin, Adele. 1991. Biblical Poetry through Medieval Jewish Eyes.
Bloomington: Indiana University Press.

Bland, Kalman. 1990. ‘Issues in Sixteenth-Century Jewish Exege-
sis’. In The Bible in the Sixteenth Century, edited by David
Steinmetz, 50-67. Durham, NC: Duke University Press.

Boyarin, Daniel. 2009. Socrates and the Fat Rabbis. Chicago: Uni-
versity of Chicago Press.

Bregman, Marc. 2003. The Tanhuma-Yelammedenu Literature. Pis-

cataway, NJ: Gorgias Press. [Hebrew]

366 Williams

Caspi, Ephraim. 2015. ‘Meqorot Darshaniyim la-Otiyot ha-
Metuyagot ve-ha-Meshunot ba-Tora’. Hitzei Giborim 8: 403-
46.

Clenman, Laliv. 2014. ‘The Fire and the Flesh: Self-Destruction
and the Male Rabbinic Body’. In The Body in Biblical, Chris-
tian and Jewish Texts, edited by Joan Taylor, 210-25. Lon-
don: Bloomsbury.

Cohen, Menahem (ed.). 1996. Mikra’ot Gedolot ‘Haketer’: Isaiah.
Ramat Gan: Bar-Ilan University Press. [Hebrew]

Cohen, Mordechai. 2020. The Rule of Peshat: Jewish Constructions
of the Plain Sense of Scripture and Their Christian and Muslim
Contexts, 900-1270. Philadelphia: University of Pennsylva-

nia Press.

. 2021. Rashi, Biblical Interpretation, and Latin Learning in
Medieval Europe: A New Perspective on an Exegetical Revolu-
tion. Cambridge: Cambridge University Press.

Cohen, Tamir. 1997-1998. ‘‘Al Ta‘am ve-‘al Ri"k’. MA thesis, Yaa-
kov Herzog Teachers’ Institute.

Dan, Joseph. 1968. The Esoteric Theology of Ashkenazi Hasidism.
Jerusalem: Bialik Institute. [Hebrew]

Dotan, Aron. (ed.). 1967. The Diqduge ha-Te‘amim of Aharon ben
Moshe ben Asher. Jerusalem: The Academy of the Hebrew
Language. [Hebrew]

. 2005. ‘Paseq in Antiquity’. In Samaritan, Hebrew and Ar-
amaic Studies: Presented to Professor Abraham Tal, edited by
Moshe Bar-Asher and Moshe Florentin, 121-33. Jerusalem:
Bialik Institute. [Hebrew]

“Some Fanciful Midrash Explanation” 367

. 2009. ‘Masora’s Contribution to Biblical Studies—Re-

vival of an Ancient Tool’. In Congress Volume Ljubljana
2007, edited by André Lemaire, 57-69. Leiden: Brill.
Dweck, Yaacob. 2010. ‘Editing Safed: The Career of Isaac Ger-
shon’. Jewish Studies Quarterly 17: 44-55.
. 2011. The Scandal of Kabbalah. Princeton, NJ: Princeton

University Press.

Encyclopedia Talmudica: A Digest of Halachic Literature and Jewish
Law from the Tannaitic Period to the Present time, Alphabeti-
cally Arranged. Edited by Meir Bar-Ilan and Shlomo Josef
Zevin, translated by Isidore Elstein and Harry Freedman. 6
vols. Jerusalem: Talmudic Encyclopedia Institute, 1969-.

Englander, Henry. 1939. ‘Grammatical Elements and Terminol-
ogy in Rashi’. Hebrew Union College Annual 14: 387-429.

. 1942-1943. ‘A Commentary on Rashi’s Grammatical

Comments’. Hebrew Union College Annual 17: 427-98.
Fine, Lawrence. 2003. Physician of the Soul, Healer of the Cosmos:
Isaac Luria and His Kabbalistic Fellowship. Stanford, CA:
Stanford University Press.
Fishbane, Michael. 1994. The Kiss of God: Spiritual and Mystical
Death in Judaism. Seattle: University of Washington Press.
. 2013. ‘Midrash and the Meaning of Scripture’. In Midrash
Unbound: Transformations and Innovations, edited by Mi-
chael Fishbane and Joanna Weinberg, 13-24. Oxford:

Littman.

Fredman, Lisa (ed.). 2019. Rashi’s Commentary on the Book of
Proverbs: Edited with Introduction and Notes. Jerusalem:
World Union of Jewish Studies. [Hebrew]

368 Williams

Freedman, David Noel (ed.). 1998. The Leningrad Codex: A Fac-
simile Edition. Grand Rapids, MI: Wm. B. Eerdmans.

Friedeman, Caleb et al. (eds). 2021. A Scripture Index to Rabbinic
Literature. Peabody, MA: Hendrickson.

Gelles, Benjamin. 1981. Peshat and Derash in the Exegesis of Rashi.
Leiden: Brill.

Gellis, Jacob. 1982-2014. Sefer Tosafot ha-Shalem. 17 vols.; Jeru-
salem: Mifal Tosafot Hashalem.

Ginsburg, Christian (ed.). 1926. The Writings: Diligently Revised
According to the Massorah and the Early Editions. London:
British and Foreign Bible Society.

Goren, Zekhariah. 1989. ‘Al ha-Mashma‘ut ha-Parshanit shel
Ta‘am “ha-shalshelet”’. Tura: 83-92.

. 1995. Ta‘ame ha-Migqra ke-Farshanut. Tel-Aviv: Hakibbutz

Hameuchad.

Grossman, Avraham. 2012. Rashi. Translated by Joel Linsider.
Oxford: Littman.

. 2017. ‘Rash"i ke-Darshan Megori’. In Transformations in
Medieval Jewish Society, 386-407. Jerusalem: Bialik Pub-
lishing.

. 2021. Rashi and the Jewish-Christian Polemic. Ramat Gan:

Bar Ilan University. [Hebrew]

Hallamish, Moshe. 1999. An Introduction to the Kabbalah. Trans-
lated by Ruth Bar-Ilan and Ora Wiskind-Elper, 281-309. Al-
bany: State University of New York Press.

Heinemann, Isaac. 1970. Darkhe ha-’Aggada. Jerusalem: Magnes

Press.

“Some Fanciful Midrash Explanation” 369

Herring, Basil. 1982. Joseph Ibn Kaspi’s Gevia‘ Kesef: A Study in
Medieval Jewish Philosophic Bible Commentary. New York:
KTAV.

Himmelfarb, Lea. 2004. ‘On One Masorah in Rashi’s Biblical
Commentary’. Sefarad 64: 75-94. [Hebrew]

. 2005. ‘On Rashi’s Use of the Masorah Notes in his Com-
mentary to the Bible’. Shnaton: An Annual for Biblical and
Ancient Near Eastern Studies 15: 174-78. [Hebrew]

Hyman, Arthur (ed.). 1973. Yalqut Shim‘oni ‘al ha-Torah le-Rab-
benu Shim‘on ha-Darshan: Sefer Bereshit. Jerusalem: Mossad
Harav Kook.

Ibn Ezra, Abraham. 1977a. Perushe ha-Torah. Edited by Asher
Weiser. 3 vols. Jerusalem: Mossad Harav Kook.

. 1977b. Sefer Sahot. Edited by Carlos del Valle Rodriguez.
Salamanca: Universidad Pontifica.

Ibn Parhon, Solomon. 1844. Lexicon Hebraicum. Edited by Salo-
mon Stern. Pressburg: Shmid. [Hebrew]

Idel, Moshe. 1983. ““We Have No Kabbalistic Tradition on This”’.
In Rabbi Moses Nahmanides: Explorations in His Religious and
Literary Virtuosity, edited by Isadore Twersky, 51-73. Cam-
bridge, MA: Harvard University Press.

. 1988. The Mystical Experience in Abraham Abulafia. Al-

bany: State University of New York Press.

Ilan, Tal. 2017. Massekhet Hullin: Text, Translation, and Commen-
tary. Tiibingen: Mohr Siebeck.

Jacobs, Louis. 2004. Jewish Preaching: Homilies and Sermons. Lon-
don: Vallentine Mitchell.

370 Williams

Kahana, Menahem (ed.). 2011-2015. Sifre on Numbers: An Anno-
tated Edition. 4 vols. Jerusalem: Magnes Press. [Hebrew]

Kamin, Sarah. 1980. ‘Rashi’s Exegetical Categorization with Re-
spect to the Distinction between Peshat and Derash’. Im-
manuel 11: 16-32.

. 1986. Rashi’s Exegetical Categorization in Respect to the Dis-

tinction between Peshat and Derash. Jerusalem: Magnes
Press. [Hebrew]

Kanarfogel, Ephraim. 1986. ‘The ‘Aliyah of “Three Hundred Rab-
bis” in 1211: Tosafist Attitudes toward Settling in the Land
of Israel’. Jewish Quarterly Review 76: 191-215.

Khan, Geoffrey. 2013. A Short Introduction to the Tiberian Maso-
retic Bible and Its Reading Tradition. 2nd edition. Piscataway,
NJ: Gorgias Press.

Kogut, Simcha. 1994. Correlations between Biblical Accentuation
and Traditional Jewish Exegesis: Linguistic and Contextual
Studies. Jerusalem: Magnes Press. [Hebrew]

Konyevsky, Hayim (ed.). 1978-1980. Rokeach: A Commentary on
the Bible by Rabbi Elazar of Worms. 3 vols. New York: Bnei
Brak. [Hebrew]

Kugel, James. 1990. In Potiphar’s House: The Interpretive Life of
Biblical Texts. Cambridge, MA: Harvard University Press.

Lange, Itzhak (ed.). 1980. Ta‘ame Masoret ha-Miqra: Le-Rabi Ye-
huda he-Hasid. Jerusalem: Ben-Arzah.

Last, Isaac (ed.). 1905. Joseph ibn Caspi’s Mishneh Kesef. 2 vols.

Cracow: Fischer. [Hebrew]

“Some Fanciful Midrash Explanation” 371

Lawee, Eric. 2008. ‘Isaac Abarbanel: From Medieval to Renais-
sance Jewish Biblical Scholarship’. In Hebrew Bible/Old Tes-
tament: The History of Its Interpretation. Volume II: From the
Renaissance to the Enlightenment, edited by Magne Sebg,
195-99. Gottingen: Vandenhoeck & Ruprecht.

. 2019. Rashi’s Commentary on the Torah: Canonization and

Resistance in the Reception of a Jewish Classic. Oxford: Ox-
ford University Press.

Levinson, Joshua. 1997. ‘An-Other Woman: Joseph and Poti-
phar’s Wife—Staging the Body Politic’. Jewish Quarterly Re-
view 87: 269-310.

Liebermann, Saul. 1962. Hellenism in Jewish Palestine. 2nd edi-
tion. New York: Jewish Theological Seminary of America.

Loewinger, David. 1960. ‘The Aleppo Codex and the Ben Asher
Tradition’. Textus 1: 59-111.

. 1972. Review of G. Weil, Massorah Gedolah: iuxta codicem
Leningradensem B19a. Kiryat Sefer 47: 601-5. [Hebrew]

Magonet, Jonathan. 2004. A Rabbi Reads the Bible. 2nd edition.
London: SCM.

Margaliot, Reuben (ed.). 2007-2008. Sefer Zohar Hadash. Jerusa-
lem: Mossad Harav Kook.

Martin-Contreras, Elvira. 2003. ‘Comments on Textual Details:
Relationships between Masorah and Midrash’. Journal of
Jewish Studies 54: 62-70.

Martini, Annett (ed.). 2010. ‘Introduction’. In Yosef Gigatilla: The
Book of Punctuation, 19-162. Kabbalistic Library of Gio-

vanni Pico della Mirandola 4. Turin: Nino Aragno Editore.

372 Williams

Matt, Daniel C., Nathan Wolski, and Joel Hecker. 2004-2017. The
Zohar: Pritzker Edition. 12 vols. Stanford, CA: Stanford Uni-
versity Press.

Mesch, Barry. 1975. Studies in Joseph ibn Caspi: Fourteenth-century
Philosopher and Exegete. Leiden: Brill.

Mondschein, Aharon. 2009. ““The Masoretes Fabricated Explana-
tions for Full and Deficient Spellings”: On Abraham Ibn
Ezra’s Struggle Against the (Ab)use of Biblical Spelling as
an Exegetical Tool’. Shnaton: An Annual for Biblical and An-
cient Near Eastern Studies 19: 245-321. [Hebrew]

Mopsik, Charles. 1987. Le Zohar: Livre de Ruth. Lagrasse: Verdier.

Morgenstern, Yossi. 1994. ‘Ha-Shalshelet ke-Ta‘am Mefaresh’.
Tura 3: 251-54.

Norzi, Jedidiah. 2005-2006. Minhat Shay on the Torah. Edited by
Zvi Betser. Jerusalem: World Union of Jewish Studies. [He-
brew]

Ofer, Yosef. 2019. The Masora on Scripture and its Methods. Berlin:
De Gruyter.

Penkower, Jordan. 1982. ‘Jacob ben Hayyim and the Rise of the
Biblia Rabbinica’. PhD dissertation, The Hebrew University
of Jerusalem. [Hebrew]

. 2010. The Dates of Composition of the Zohar and the Book
Bahir. Los Angeles: Cherub Press. [Hebrew]

Price, James. 1996. Concordance of the Hebrew Accents in the He-
brew Bible. 5 vols. Lewiston, NY: Edwin Mellen.

Razhabi, Yitzhak. 1978. ‘Irregular Letters in the Torah’. In The
Script of the Torah and Its Characters, edited by Menahem

“Some Fanciful Midrash Explanation” 373

Kasher, 1-234. Torah Shelemah 29. Jerusalem: Beth Torah
Shelemah. [Hebrew]

Reich, Nurit. 2006. ‘The Names of the Accent Shalshelet’. Massorot
13-14: 203-24. [Hebrew]

Revell, E. J. 2000. ‘The Interpretative Value of the Massoretic
Punctuation’. In Hebrew Bible/Old Testament: The History of
Its Interpretation. Volume I/2: The Middle Ages, edited by
Magne Sebg, 64-73. Gottingen: Vandenhoeck & Ruprecht.

Rock, Avigail. 2007. ‘R. Yosef Ibn Kaspi’s Biblical Exegesis: Exe-
getical Methodology and a Critical Annotated Edition of
Mazref la-Kesef on Genesis. PhD dissertation, Bar-Ilan Uni-
versity. [Hebrew]

Rosen-Zvi, Ishay. 2011. Demonic Desires: Yetzer Hara and the
Problem of Evil in Late Antiquity. Philadelphia: University of
Pennsylvania Press.

Rubin, Aaron. 2011. ‘Samuel Archivolti and the Antiquity of the
Hebrew Pointing’. Jewish Quarterly Review 101: 233-43.

Sacks, Jonathan. 2009. Covenant & Conversation: A Weekly Read-
ing of the Jewish Bible. Genesis: The Book of Beginnings. New
Milford, CT: Maggid Books.

Sackson, Adrian. 2017. Joseph ibn Kaspi: Portrait of a Hebrew Phi-
losopher in Medieval Provence. Leiden: Brill.

Safrai, Ze’ev. 2018. Seeking out the Land: Land of Israel Traditions
in Ancient Jewish, Christian and Samaritan Literature (200
BCE — 400 CE). Leiden: Brill.

Saperstein, Marc. 2014a. ‘Cultural Juxtapositions: Problemati-
zing Scripture in Late Medieval Jewish and Christian Exe-

gesis’. In Leadership and Conflict: Tensions in Medieval and

374 Williams

Early Modern Jewish History and Culture, 113-39. Oxford:
Littman.

. 2014b. ‘The Land of Israel in Pre-Modern Jewish
Thought: A History of Two Rabbinic Statements’. In Lead-

ership and Conflict: Tensions in Medieval and Early Modern
Jewish History and Culture, 271-90. Oxford: Littman.

Schiffer, Ira Jeffrey. 1977. ‘The Men of the Great Assembly’. In
Persons and Institutions in Early Rabbinic Judaism, edited by
William S. Green, 237-76. Missoula, MT: Scholars Press.

Shalem, Shim‘on. 1965-1966. Rabbi Moshe Alshekh. Jerusalem:
Ben Zvi-Institute.

Shereshevsky, Esra. 1972. ‘The Accents in Rashi’s Commentary’.
Jewish Quarterly Review 42: 277-87.

. 1982. Rashi: The Man and his World. New York: Sepher-
Hermon Press.

Shinan, Avigdor (ed.). 1984. Midrash Shemot Rabba: Chapters I-
XIV. Jerusalem: Dvir. [Hebrew]

. 1994. ‘The Midrashic Interpretations of the Ten Dotted

Passages of the Torah’. In The Bible in the Light of its Inter-

preters, edited by Sara Japhet, 198-214. Jerusalem: Magnes
Press. [Hebrew]

Sinai, Yuval. 2007. ‘Judicial Authority in Fraudulent-Claim Cases
(din merume)’. Jewish Law Annual 17: 209-66.

Smelik, Willem. 2013. Rabbis, Language and Translation in Late
Antiquity. Cambridge: Cambridge University Press.

Talmage, Frank. 1999 (1986). ‘Apples of Gold: The Inner Mean-
ing of Sacred Texts in Judaism’. In Apples of Gold in Settings

of Silver: Studies in Medieval Jewish Exegesis and Polemics,

“Some Fanciful Midrash Explanation” 375

edited by Barry Walfish, 108-50. Papers in Medieval Stud-
ies 14. Toronto: Pontifical Institute of Mediaeval Studies.
Originally published in Jewish Spirituality I: From the Bible
through the Middle Ages, edited by Arthur Green, 313-55.
New York: Crossroad.

Ta-Shma, Israel. 2005. ‘Midrash Lekah-Tov: Its Background and
Character’. In Studies in Medieval Rabbinic Literature: 3. Italy
& Byzantium, 96-114. Jerusalem: Bialik Institute. [Hebrew]

Theodor, Julius, and Chanoch Albeck (eds.). 1903-1936. Midrash
Bereshit Rabba. 3 vols. Berlin: Ittskovski.

Touitou, Elazar. 2000. ‘Darko shel Rashi be-Shimusho be-
Midreshe Haz"al: ‘Iyun be-Ferush Rashi li-Shemot 1:8-22’.
Talele ?Orot 9: 51-78.

Twersky, Isadore. 1979. ‘Joseph ibn Kaspi: Portrait of a Medieval
Jewish Intellectual’. In Studies in Medieval Jewish History
and Literature, 3 vols., edited by Isadore Twersky, I:231-57.
Cambridge, MA: Harvard University Press.

Van der Heide, Albert. 1983. ‘PARDES: Methodological Reflec-
tions on the Theory of the Four Senses’. Journal of Jewish
Studies 34: 147-59.

Walfish, Barry. 1993. Esther in Medieval Garb. Albany: State Uni-
versity of New York Press.

Werblowsky, R. J. Zwi. 1997. Joseph Karo: Lawyer and Mystic.
Philadelphia: Jewish Publication Society of America.
Wickes, William. 1887. A Treatise on the Accentuation of the
Twenty-One So-Called Prose Books of the Old Testament. Ox-

ford: Clarendon.

376 Williams

Williams, Benjamin. 2015. ‘Doubting Abraham Doubting God:
The Call of Abraham in the Or ha-Sekhel’. Melilah 12: 31-
42.

. 2016. Commentary on Midrash Rabba in the Sixteenth Cen-
tury: The Or ha-Sekhel of Abraham ben Asher. Oxford: Ox-
ford University Press.

Wolfson, Elliot. 1988-1989. ‘Biblical Accentuation in a Mystical
Key: Kabbalistic Interpretation of the Te’amim: Part 1’. Jour-
nal of Jewish Music and Liturgy 11: 1-16.

. 1989. ‘By Way of Truth: Aspects of Nahmanides’ Kabba-

listic Hermeneutic’. AJS Review 14: 103-78.

. 1989-1990. ‘Biblical Accentuation in a Mystical Key:

Kabbalistic Interpretation of the Te’amim: Part 2’. Journal
of Jewish Music and Liturgy 12: 1-13.
. 1993. ‘Beautiful Maiden Without Eyes: Peshat and Sod in

Zohardic Hermeneutics’. In The Midrashic Imagination: Jew-

ish Exegesis, Thought, and History, edited by Michael
Fishbane, 155-203. Albany: State University of New York
Press.

Wright, Charles. 1864. The Book of Ruth in Hebrew. London:
Williams & Norgate.

Yeivin, Israel. 1980. Introduction to the Tiberian Masorah.
Translated and edited by E. J. Revell. Missoula, MT:
Scholars Press.

Yisraeli, Oded. 2006. ‘The Kabbalistic remez and Its Status in
Nahmanides’ Commentary on the Torah’. The Journal of
Jewish Thought and Philosophy 24: 1-30.

DOES SAADYA REFER TO THE ACCENTS
IN HIS INTRODUCTION TO THE
PENTATEUCH??*°

Joseph Habib

In his article ‘The Tension between Literal Interpretation and Ex-
egetical Freedom’, Haggai Ben-Shammai (2003, 38, n. 42) raised
the possibility that the famous medieval rabbinic scholar Saadya
Gaon (882-942) directly refers to the biblical accents and their
function of joining and separating words. The relevant passage
comes from Saadya’s introduction to his long commentary on the
Pentateuch (henceforth SIP). A thorough analysis of the passage
was beyond the scope of Ben-Shammai’s article. Saadya does not
explicitly mention the accents in SIP, but what is clear is that
Saadya attaches exegetical importance to the grouping of words
in a passage. The purpose of this article is to determine whether

or not Saadya has specifically the accents in mind.

8° The content of this article formed part of my PhD research, which
was funded in part by the Faculty of Asian and Middle Eastern Studies,
University of Cambridge, and in part by the University of Haifa’s Valler
Doctoral Fellowship. I warmly thank them for their generosity and sup-
port.

378 Habib

1.0. Saadya Gaon and the Importance of His

Testimony

If the phenomenon that Saadya describes in SIP does indeed re-
late to the accents, this is significant because (1) Saadya had di-
rect contact with the Tiberian Masoretes and (2) the question of
the original function of the biblical accents remains unanswered
(see, recently, De Hoop 2008; Shoshany 2009; Park 2014; Pitcher
2020). Saadya’s account would therefore furnish testimony rele-
vant to the function of the accents contemporary with the Tibe-
rian Masoretes.

Saadya, known in Arabic as Sa‘id ben Yiisuf al-Fayyiimi,
was born in Egypt around 882 cE (the most authoritative bio-
graphy remains Malter 1921). Throughout his life, Saadya’s crea-
tive mind and wide range of knowledge allowed him to make
foundational contributions to a number of intellectual fields, in-
cluding biblical exegesis, grammar, poetry, and halakha (Brody
2006 [Hebrew]; 2013 [English]). Saadya’s capacity as a scholar
led to his appointment as head—Gaon—of the struggling Baby-
lonian academy (n12°w” yeshiva) in Sura in 928 (Brody 1998, 237-

238).”*” His most significant contribution to biblical exegesis was

?87 Tt was once thought that it was Saadya’s involvement in the calendar
controversy that erupted in 921/2 cE between the Babylonian and Pal-
estinian yeshivot that led to his appointment as Gaon. On the basis of a
fresh examination of the sources, Stern (2019) has demonstrated that
Saadya’s role in the exchange between the two academies was marginal.

Does Saadya Refer to the Accents? 379

his translation of, and commentary on, most of the Bible, conven-
tionally known as the tafsir.?°

Before his appointment as Gaon, Saadya spent about ten
years in the city of Tiberias. For some of that time, he was the
student of one Abi Kathir Yahya ben Zakariyya, whom the Mus-
lim historian al-Mas‘tidi (893-956) describes as a “Tiberian
scribe (al-katib al-tabarani)” and as an expert in Bible transla-
tion.**° In that time, Saadya would have been exposed to a variety
of reading traditions within different masoretic circles (Dotan
1996).

?38 The appellation tafsir was not used by Saadya himself, but has be-
come accepted among scholars; I thank Ronny Vollandt for pointing this
out to me. There is not yet general consensus as to when Saadya began
this project or how it developed. See Zewi (2015, 28-29) for a recent
discussion of the different arguments.

739 See de Goeje (1894, 112-13) for the original source; see also Gil
(1992, 176-78), Polliack (1997, 11-12). The second source from which
Saadya’s time in Palestine is known is a letter he wrote to his three
students. The scenario was as follows: Saadya and one R. David were
both in Babylon. R. David received a letter from Saadya’s students, who
requested letters from the Babylonian academy regarding the calendar
controversy of 921/2 cE, in which Saadya played a small part (Stern
2019, 288; many thanks to Prof Ronny Vollandt for this reference). Puz-
zled as to why his students had not written to him instead, Saadya wrote
back to them: Any ty > ont 1D nda “tydan pox onand xb 7D KN TNA09
Sxqw” pixa ‘ty ‘I believe that you only wrote to him, rather than to me,
because you assumed that until now I was still in the land of Israel’ (text
and translation from Stern 2019, 308-9; this fragment was published
earlier by Schechter 1901, 60, fol. 1v, Ins 6-8 for the original letter
fragment; see also Brody 2013, 26 for comment on the letter).

380 Habib

Saadya’s intellectual connection with the Masoretes is evi-
dent in a number of places throughout his works. Traces of the
masora are seen in Saadya’s grammar book The Book of Elegance
of the Language of the Hebrews (Dotan 1997, 34-35). In his chapter
on vowels, Saadya clearly derives material from the masoretic
treatise now known as Okla we-’Okla and refers to it as ‘the ma-
sora’ (m70NNdX; Dotan 1997, 433). There he tells his reader to
consult ‘the masora’ and then lists words in which the final he is
not pronounced with mappigq, nearly all of which are found in a
list in °?Okla we-’Okla.*” Dotan (1997, 35-36) notes that Saadya
drew heavily from the m710AN YoTNUNP quntrese ha-masora, i.e.,
masoretic treatises, when formulating his rules for shewa.**! An-
other point of contact between Saadya and the Masoretes is re-
flected in a disagreement among them. At one point in the gram-
mar (Dotan 1997, 410), Saadya objects to a masoretic formula-
tion of the rule for fricativisation of the n"59732 letters. Saadya
states that the rule should be that this realisation depends on
whether the previous word ends in a vowel, not, as some Maso-
retes formulated the rule, on whether the previous word ends in
the letters x, 1, °, or 7 (see Ofer 2019, 234).

240 See Diaz Esteban (1975, 85-86) or the list in the treatise. The second
part of the treatise was published by Ognibeni (1995).

?41 The term quntrese ha-masora was coined by Dotan (1967, 13) to refer
to instances which represent the first attempts at formulating systematic
rules based on the masora. They were variously copied individually, as
a group, or found in the pages of other books, but never formed a stand-
alone work, such as that of Diqgduge ha-Te‘amim (Dotan 2005, 20).

Does Saadya Refer to the Accents? 381

Regarding the accents specifically, Saadya composed some
of his works in the style of Biblical Hebrew, complete with Tibe-
rian vowel and accent signs. One such work is the Hebrew intro-
duction to his dictionary of Biblical Hebrew, entitled Ha-’Egron
‘The Thesaurus’ (ed. Allony 1969). Saadya’s use of the accents
generally conforms to the rules with which modern scholarship
is familiar (Revell 1974, 125, and, more recently, Hitin-Mashiah
2011, though Revell also notes many peculiarities in Saadya’s use
of accents). It differs only in small details, which is to be ex-
pected, given the fact that Saadya lived and worked nearly three
centuries before Maimonides proclaimed the Ben Asher tradition
as authoritative (Penkower 1981; Ofer 2019, 144-46). This
would have therefore been a time when various sub-traditions of
the Tiberian Masoretic reading tradition would have existed side-
by-side. The reason given by Saadya himself for his use of accents
in his works is “so that its (i.e., the text’s) reading may be easier
and its memorisation more possible (j2Ax1 ANNA SADK I>
mvan>)” (Yeivin 1959, 48). Saadya is probably pointing here to
some kind of aural (and oral) phenomenon that he would have
expected the reader to hear in their mind’s ear (Habib 2021, 35).
It is still not clear exactly what he is referring to, but Revell
(1974, 125ff.) has argued that Saadya must have thought of the
accents as an organic part of Biblical Hebrew.

Given Saadya’s geographical, chronological, and intellec-
tual proximity to circles of the Masoretes, any comment he offers
on the accents would afford scholars valuable insight into “one
of the most neglected fields in the study of Hebrew graphemes”

(Dotan 1970, vii). If Saadya does indeed have the accents in mind

382 Habib

in SIP, then the passage should receive serious consideration in
future scholarship on the accents.

2.0. Analysis of SIP

The passage in question comes from what is conventionally
known as Saadya’s ‘long commentary’ on the Pentateuch—an ex-
egetical work that consists of translation of biblical verses em-
bedded within a ‘long commentary’.*” The passage, along with

my translation, follows.?”

Zucker (1984, 19) _Lines

70 parxrdsx popoxi The fourth part of those things that clar-
poxuds nangids a ify obscurities is the joining/ grouping
IRD NNALIN on of words. So, the commonly accepted

TNX RTTAWA sense of it (a verse) is taking [it]

242 A name for the ‘long commentary’ is not extant (Ronny Vollandt,
personal correspondence). Saadya called his translation of the Penta-
teuch devoid of commentary ikhraj ma‘ani nass al-tawrat ‘extraction
(i.e., edition) of the meanings of the text of the Torah’ (see Vollandt
2015, 82-83); this is known in scholarly parlance as Saadya’s ‘short
tafsir’. The remainder of the biblical books that Saadya translated and
commented upon each bear their own unique title consistent with the
content of the book (Vollandt 2015, 82). Saadya titles his commentary
on Isaiah, for example, kitab ’al *istilah I-’al-ta‘a, which, according to
Ben-Shammai (1991), should be understood as ‘The Book of the Endeav-
our towards Improvement of Worship’.

?43 Tn all translations of Judaeo-Arabic texts that follow, square brackets
‘[]’ indicate an editorial addition to make the text more readable; round
brackets ‘()’ indicate an editorial comment for clarification and biblical
references.

Does Saadya Refer to the Accents? 383

182 TNT ANY nN]
*) nov manda 52 n>
*yn Spin AR poR
> TNNI

pip xd poands
aoy xd ToT NAP.
pa voandsr 7
Aprandsy nxavad>x
syoan a dipx “na
Synn pty Daw
pa a2 prpds yx
DYN IX&

Atn Sy ponds
7N1.ATN dy [ANd ]on
dndndx 75 ixnnx
sunpnds wend

DInwA 17 ND PID’ TR
pan 17 NA ADP IX
757 TRA ARDS 78

xn2

Tn ndip ox pry
WN TNX Wr ONT
nn yn aw nyt
syyndx opno?
porn 7p ToT
Aya EK m7
DIPWA Dy

Ty WII WAND|D
opwa

according to the way it is arranged to-
gether. This because all the words that
are joined together in the reading
should considered to have one meaning
in

the interpretation, and they should not
be divided. Similarly, words that are
separated should not be joined together
in interpretation,

just as I will say in the interpretation of
®Ovian | oy oq (Isa. 6.2) that there
is a big difference between whether two
words

are joined together or one word stands
apart from the others. If there is need to
deviate [from the accepted meaning] so
that the sense [units] are broken apart
(lit., harmed)—

[ie.,] if that which is joined is sepa-
rated, or that which is separated is
joined in the reading—that is possible,
just as

the phrase F770 DTNA 17 ODN AIT | WAN
yr aiv nyt) 13 Tnx (Gen. 3.22) is sep-
arated and joined in order that the in-
terpretation be correct.

This is also the case in the joining of
PTA ADD OPW DVI AVIS TAH
(Exod. 25.34; 37.20). [The word]

o-ipwn is intended to be

384 Habib

in the first [part], not within the sec-

ry xd Sindee ry pw eet
: : ond, as well as that which is similar to
yori RDDUNTDR = 25
i that division (i.e., the other verses like
popdxi 757.

this one. [Exod. 25.33; 37.19, 20]).

The text that precedes this passage is missing from the man-
uscript and, therefore, the broader context of the passage is una-
vailable. The opening lines make it clear enough, though, that it
may be interpreted as a self-contained paragraph. Evidently this
passage constitutes but one item in a list which deals with how
to “clarify obscurities (Poxudx nxangindsx)” (In. 17) one encoun-
ters in the Biblical text. In this analysis, I will focus on three key
terms which must be properly understood in order to determine

whether or not Saadya is referring in this passage to the accents.

2.1. Dammat

The fourth item in this list of the principles that “clarify obscuri-
ties” Saadya designates with the term nxnv>x al-dammait (In. 17).
I have translated this term ‘joining/grouping of words’. One
method employed by Saadya to clarify obscurities is appeal to al-
dammat—the joining and grouping of words. The grouping of
words in the context of disambiguating the sense of a passage
would correspond to the accents’ function of joining and separat-
ing words. This may indeed be what Saadya intended. Though
Ben-Shammai (2003, n. 41) states that he has not come across
the word dammat elsewhere in Saadya’s corpus, I have found one

further instance of Saadya’s use of this word with what seems like

Does Saadya Refer to the Accents? 385

a similar meaning, on the basis of which the sense of the term

may be clarified.

Zucker (1984, 103) Lines

mynd) 7s xp
pya sy pya binds
you Pia ons ipa
noIbx

ipa pyr dx R&va
qOn Paw TAWRI
dio prwids ody
TANPR NN

It is said that in its (i.e., Scripture’s)
grouping the men together in the phrase
JAX PID AMX ‘You and your sons with
you”* (Gen. 6.18) and the grouping of
the women

together in the phrase P32- wn FAWN) ‘And
your wife and your son’s wife’ (Gen. 6.18)
forbids sexual relations for the duration of

which they are

p. 104

Op fim] mands 7
Day NO ANP.
say ANNSON vIn
pxp jndndx

jpa andr >s dr[dx]
WI PID TNWRI ANN
7,12

in the ark. This statement is likely to be
correct (lit. close [to the correct interpre-
tation]). That which is opposite this
grouping of words in [the account of]
their exiting [the ark] strengthens this [in-
terpretation]. So, [there], it joins

the husband to his wife in the phrase 7mx
PII-VH PII JAW ‘You and your wife
and your sons and your sons’ wives’ (Gen.
8.16).

Saadya argues here that one may interpret the actual

grouping of words used in Gen. 6.18 as a prohibition which God

gave to Noah to abstain from sexual relations while on the ark.

To support this claim Saadya points to the account of their exiting

44 In Gen. 6.18 the word 7nx is at the end of the verse, not after 7nx

Tad.

386 Habib

the ark (Gen. 8.16) and says that the dammat there is the opposite
(o2y; p. 104, In. 1)—Noah is grouped with his wife (qnwxi nAY)
and his sons are grouped with their wives (7?12-°WH 77321). Since
they are leaving the ark, there is no longer concern for their sex-
ual relations while inside. In this context, the word dammat ob-
viously refers to the grouping of the items within both lists. We
therefore see similarity in this use of the term dammat with what
is found in SIP, in that both indicate the grouping of words.

The use of dammat in the commentary on the Genesis pas-
sages does not reflect the grouping of words according to the ac-
cents. I will argue below that this is actually a crucial clue for the
question addressed in this essay. First, however, it should be no-
ticed that in both places Saadya’s interpretation is based on a
2+ 2 division of the list, whereas the accents divide both lists into
14+3.

Gen. 6.18: Entering Gen. 8.16: Exiting

TAR TON TN] ANN
Saadya
TAN TON} TION TAN
TAX TAX
Accents

TAUNTS TIB Paw Ta Tw)

In both verses, the list falls within the domain of the accent silluq.
The accents which divide silluq’s clause are zagef (main division)
and tifha (minor division). In both lists the sequence of accents is
the same—the zagef (gadol) accentuates the first item of the list
(nx in both cases), leaving the next three items to be grouped

together and terminating in tifha. If, for Saadya, the dammat were

Does Saadya Refer to the Accents? 387

based on the accents, it would create a difficulty for his commen-
tary on Gen. 6.18. There, the accents group Noah’s sons (77311)
with the women (7732°wN JAWR).745

It appears that the dammat Saadya refers to reflects rabbinic
tradition, rather than the accents (the reason for which will be

made clear below). The Babylonian Talmud states:

TAS PIAWH TAVRIPIa Ans Aan Ox nga nat no0x27 rin
7) ANP 8) TSX PI2 WI PId TOW) nS AANA JA && and)
NOAN WW? NOXIW Asx

From where do we know that they were prohibited [from
sexual relations on the ark]? From that which is written,
“And you will come to the ark—you, and your sons, and
your wife, and your sons’ wives with you” (Gen. 6.18). And
it is [also] written: “Go out from the ark—you, and your
wife, and your sons and your sons’ wives with you” (Gen.
8.16). R. Johanan said, “From here they said that they
were prohibited from sexual relations on the ark” (b. San-
hedrin 108b.14).

This interpretation is paralleled in the Jerusalem Talmud:

mand inoa3 ni pny’ 2772 bxinw owl pin TD AAT
WH TNWRI PID ANS ann bx nal "n AON wnAwn Dd AON
ANS AINA 7k’ vO" AwAN ww 1 AINA InNK’A TNX PIA

TNX Pawn PII Tawi
R. Judah b. Pazi, R. Hanin on behalf of R. Samuel b. Rav
Issac [say], “When Noah entered the ark sexual relations
were forbidden to him. [For,] what is the meaning of, “And
you will enter the ark—you, and your sons, and your wife

4° The word 7°13) is out of necessity accented with the minor disjunctive
tebir, because tifha’s clause can only contain three words (see Wickes
1887, 89; cf. Breuer 1989, 45).

388 Habib

and your sons’ wives with you” (Gen. 6.18)? When he

(Noah) came out, he was permitted to have sexual rela-

tions. [For,] what is the meaning of, “Go out from the

ark—you, and your wife, and your sons and your sons’

wives” (Gen. 8.16)? (y. Ta‘anit 7a.1)

This parallel suggests that, even if Saadya’s use of dammat
in SIP above refers to the grouping of words by the accents, here
the term was used to mean the grouping of words according to
the interpretation of the Oral Law. This apparent contradiction
in usage by Saadya is elucidated by a passage in his Kitab al-
?amanat wal-’itqadat ‘The Book of Beliefs and Opinions’. At the
beginning of his chapter on resurrection, Saadya explains that
any given scripture may be interpreted in a way other than its
apparent (zahir) sense for one of four reasons. As his fourth rea-
son, he states: “Anything to which tradition applies a condition,
we will interpret it in agreement with the reliable tradition ( xm
APTRYDR ANANDOR pO’ x yWoAn mo|a md>y AvNwa aAXRNMNdS na nRi)”
(Qafih 1969, 220, Ins 6-8). The case of Gen. 6.18 is one where
this exception clearly applies. This raises an important point—the
exception highlights the rule. If tradition allows for an interpreta-
tion of Gen. 6.18 which departs from the ‘apparent’ (zahir) group-
ing of words (dammadt), one must ask what the ‘plain’ or ‘appar-
ent’ (gahir) grouping of words is from which the tradition’s inter-
pretation departs? What is it that governs this grouping of words?
The accents are one obvious possibility. But, again, Saadya never
explicitly says this.

In sum, in his commentary on Gen. 6.18 Saadya used the

term dammat to refer to groupings of words that were not re-

Does Saadya Refer to the Accents? 389

flected by the accents. This grouping was based rather on rab-
binic tradition. The question arises whether this was the principle
of grouping that Saadya referred to also in SIP cited above. This
may have been the case. Another possibility, however, is that
Saadya used the term by default to refer to groupings reflected
by the accents where no other factors were at play, but used it to
refer to other types groupings when these were sanctioned by

rabbinic tradition.

2.2. Mashur

Returning to the passage in SIP, the next idea which may point
to Saadya’s use of the term dammat as referring to the accents is
the word "1nwn mashir ‘commonly accepted’ (In. 17). Elsewhere
in his introduction to Genesis, Saadya uses the term mashir to
clarify the meaning of the word zahir ‘apparent’. He says, 7px
Senynoxdsy 7nl>x1 ans Sax pa xn a tinwnds ‘[By zahir] I mean that
which is commonly known (mashir) among native speakers of
the language, as well as that which is used frequently’ (Zucker
1984, 18, Ins 1-2). Ben-Shammai (2003, 37) highlights the two-
fold nature of Saadya’s description of mashir in another pas-
sage.”*° On the one hand, the frequency with which words occur
in a written text is measured by their distribution throughout that
text. On the other, Saadya’s equation of mashur with zahir points
to the fact that features which later readers of a text can declare
mashur, based on a measurement of their distribution, must re-

flect what the zahir ‘plain sense’ of that feature was to the lan-

246 See also Ben-Shammai (1991, 380) for a brief discussion of mashir.

390 Habib

guage’s original users. It seems to me that Ben-Shammai is draw-
ing attention to the fact that for Saadya, mashir included the idea
of what something would have meant to the native speakers of
Biblical Hebrew. This meaning becomes ‘commonly known’
among the readers of Biblical Hebrew only on the basis of its
distribution (and, necessarily in this case, its frequency) through-
out the Biblical text.

In SIP, Saadya says that the mashir of a given verse includes
its interpretation according to the ‘grouping’ of its words (dam-
matihi; Ins 17-18). If we apply the aforementioned definition of
mashur here, the implication is that adhering to the dammat of a
passage of Scripture yields its original understanding/interpreta-
tion (mashir). This raises an important question—how would
Saadya have known the divisions of a passage that reflect a so-
called ‘native speaker’s’ original understanding of Biblical He-
brew? One clear possibility is the prosodic divisions reflected by
the accents, if these were considered by Saadya to be an integral
component of the spoken language. If, against Ben-Shammai’s
suggestion, we were to understand mashur with the meaning of
‘commonly accepted’, this could be equated with the default in-

terpretation of groupings of words on the basis of the accents.

2.3. Al-qgird’a

The final important term for understanding SIP is AXIP>RX al-qird’a
‘the reading’ (In. 18). After the statement about mashur, Saadya
elaborates and explains that the purpose of the dammat ‘grouping

of words’ is so that “words which are joined together in the read-

Does Saadya Refer to the Accents? 391

ing (Ax pdx; In. 18) would have one meaning in the interpreta-
tion (Woandx; In. 19).74”7 The challenge here lies in determining
whether the term al-qird’a ‘reading’ is being used in a specific or
general sense. The former would most likely denote the oral/mu-
sical recitation of the accents, while the latter would not. The
determination must be made on the basis of Saadya’s use of this
word in other places.

Saadya’s use of al-qird’a in his introduction to the Psalms
strongly suggests that the intention is indeed oral recitation.7**
There Saadya explains that “the people would sing praises to God
with it (the Psalms) according to five stipulations (xix. nip>x
vonw nond >y otp>x °p na nnao)” (Qafih 1966, 30, Ins 9-10). Alt-
hough the text of the third stipulation is fragmentary, enough has
been preserved to understand that Saadya is arguing that the in-
struments to be used in the accompaniment to any particular
psalm must be chosen on the basis of the psalm’s superscription.
For example, any superscription which contains 40x) ‘for/by
Asaph’ (e.g., Ps. 53.1; 74.1; 75.1; etc.) must be accompanied by
a cymbal, due to the Bible’s association of the two in the verse
ynwn onyaa yox) ‘And Asaph sounds the cymbals’ (1 Chron.
16.5). In the final part of this stipulation Saadya states:

247 “The accents are probably indicated here by the Arabic term al-
qird‘a, i.e., recital” (Ben-Shammai 2003, n. 42).

?48 On Saadya’s introduction to the Psalms and for a translation of it see
Soklow (1984).

392 Habib

AR pos dap ~>y na ps Nox ANDDR ow Dy ORD TR TH RD TARA
maxdndxi

When it (the superscription) is silent (i.e., when no instru-

ment can be ascertained from it), it is not possible for it

(the Psalm) to be said with anything at all, except that it

be (said) in the manner of reading (al-qirda‘a) or recitation

(al-tilawa). (Qafih 1966, 32, Ins 12-14)?”
In other words, if no instrument is mentioned or hinted at in the
superscription, the psalm should not be accompanied by an in-
strument, but instead either be ‘read’ or ‘recited’. The significance
of this passage for the meaning of the word al-qird’a lies in its
relationship to the other manner of reading the text, indicated by
mxdnbx al-tilawa. That is, it appears that these two words reflect
alternative manners of reading the biblical text—in this case the
Psalms. Earlier in the same introduction, in the section which in-
troduces the first stipulation, this distinction is clarified:

mdip 7s on onary odds qn nip> payn int 59 78 NTN DNdxa
pa anixdn xdx dip? ps ond nb xin

The first among them (the stipulations) is that every psalm
which is ascribed to the Levites requires them to say it (the
psalm). Anyone besides them is not allowed to say it unless
(they say) its recitation only. (Qafih 1966, 30, Ins 10-13;
emphasis mine: JH).”°°

49 Cf. Soklow’s (1984, 163) idiomatic translation: ‘...a mute Psalm can-
not be put to any tune at all, it can only be read or recited.’

250 Cf. Soklow’s (1984, 158-59) translation: ‘The first of these is that
every psalm is addressed to a group of Levites who were required to
chant it. No one other than they could chant it, only recite it.’

Does Saadya Refer to the Accents? 393

As Qafih explains in his apparatus to this text, Saadya is stipulat-
ing that a psalm which is associated with the Levites can be mu-
sically chanted only by a Levite. If anyone else recites the psalm,
they must simply read it without chanting. The term al-tilawa
‘recitation’ here must therefore refer to reading without chanting.
Returning to the text of the third stipulation, it stands to reason
that this same term al-tiladwa ‘recitation’ must be juxtaposed with
a term meaning to read with musical chanting—al-qird’a ‘read-
ing’. For Saadya, then, at least some instances of al-qird’a ‘read-
ing’ refer specifically to the act of chanting the text with melo-
dies.

These considerations suggest that it is reasonable to under-
stand Saadya’s use of al-qird’a in SIP (Zucker 1984, 19, In. 18) in
specific reference to the musical cantillation of a text. The con-
text can certainly accommodate this interpretation. The joining
and separation of words is indeed one function of the accents. As
we have seen, the meaning of al-qird’a includes a musical ele-
ment—a feature that also clearly applies to the cantillation of the

accents.

2.2. Analysis of Biblical Passages from SIP

We have seen that the terms dammat ‘grouping of words’, mashur
‘commonly known’, and al-qird’a ‘reading’ can all have nuanced
meanings in the writings of Saadya. If these meanings are applied
to the passage under consideration, it seems that Saadya is saying
something to the effect of the following: one way of resolving an

obscurity in the text is by considering the grouping of its words

394 Habib

(dammat). This arrangement reflects the way in which the origi-
nal users of Biblical Hebrew (or those in the Rabbanite commu-
nity) would have uttered (or understood) the obscure passages
(=mashur). Specifically, this arrangement refers to the way in
which words are joined and separated in the oral performance
(al-qird‘a) of the text. The accents may be the common denomi-
nator of all these ideas, and so it is highly likely that this passage
refers to them despite the fact that Saadya does not mention them
explicitly.

Saadya tells his readers that he will illustrate this principle
in his commentary on Isa. 6.2. Unfortunately, that specific por-
tion of his commentary on Isaiah has not survived. But other
statements by Saadya on this verse do survive. In the very same
passage, Saadya specifies that “there is a great difference be-
tween whether two words are joined together or one word stands
apart from the others” (SIP Ins 20-21). A discussion of Isa. 6.2
also appears in Saadya’s commentary on Prov. 25.11. Below is
Saadya’s translation of Isa. 6.2 followed by the discussion of this

verse in his commentary on Proverbs.*°"

(1) Apa | Awa TDN? DrDID WY DrBID WY Tova | OIAy DTW
pip AW PZ AED oN rip
‘The Seraphim standing above were His. Each had six
wings. With two it (i.e., one of the Seraphim) would cover
up its face, and with two it would cover up its feet, and

with two it would fly.’ (Isa. 6.2)

51 All of my translations of biblical passages reflect the division of the
accents unless otherwise stated or shown.

Does Saadya Refer to the Accents?

395

Ratzaby (1993, 14) Lines
sere Aen a) : And [belonging] to Him were
angels standing in
nxt 999 ANS AND yo the heights. Each one had six
No pana " wings. With two he would hide
ieee ema an his face and with two he would
yo pina 5 hide his feet and with two he
would fly to and fro.
Qafih (1976, 199)
[The second consideration of
arranging words according to
oxdodx ya ayrnas 59 o&n yh eva eue Rea
viby 23 ranged with regard to phrasing
is that] those groups of words
which are required to be joined
together
*aNpADN Tahn xa op any by the meaning are joined to-
*y nia IND me gether. If they are joined
differently, it (the different
NTTDDN TITANIA ~—-25——joining) distorts it (the mean-
ing).
177 31~—s Similar is the
1 Dynan ovtay osnw ip phrase i> Syn | oY DS Ww. If
fon ogee 9? he/it combines three
AYANTDNI TIN YY nxn words together and the fourth
182. NTN YY es by itself, then the
29 nNDA TD ADDN ys ipo a phrase becomes, ‘Behold, God

181 ,RODON

has angels in the heavens.’ And if

396 Habib

p. 200
PNINN 77 Yy pNanK og 17 he/it joins two words together
at 1
v and two words together,
Oxp on omy oarw Oxpa so that it says | o-Jny DSW and
5 7
Nx 19 Dynan then it says i> Svinn, then this

ts : [interpretation] becomes blas-
n>xdnbds Spi tx gna. 757 B

pape pr 3 phemous, because the angels

are set above their Creator.

The context of this discussion is Saadya’s commentary on
wIDXy aT AAT oD NPawNa ant Mian ‘Like apples of gold in a
silver fixture is a word spoken in the proper way’ (Prov. 25.11).
Here, Saadya discusses at length what it means for something to
be spoken properly. One of the requirements for proper speech,
Saadya says, is that “words which are required to be joined to-
gether by the meaning (ixyn>x nahn xn) are joined together”
(Qafih 1976, 199, Ins 23-24). Failure to do so results in a distor-
tion of meaning (Ins 24-25).

In order to illustrate the aforementioned point Saadya dis-
cusses the first clause of Isa. 6.2 1) 5vian | ovTny oS. He explains
that if the words are joined 3+ 1, then the interpretation is that
“God has angels in heaven (xaodbx 7) maxon > 55x)” (Qafih
1976, 199, In. 34). That is, i ‘to him’ is the predicate of the
phrase. The subject is n’53v ‘Seraphim’ followed by an asyndetic
relative clause that modifies it—>yian | ovTny ‘(Seraphim which
are) standing above’. This is the interpretation reflected in Saad-
ya’s translation (Ratzaby 1993, 14, In. 3). Saadya then states that

arranging the clause 2+ 2, represented by the translation ‘Sera-

Does Saadya Refer to the Accents? 397

phim were standing above him’, is blasphemous, because it im-
plies that the “angels are set above their creator” (Qafih 1976,
200, Ins 2-3). While 0°53 ‘Seraphim’ serves as the grammatical
subject in either case, the predicate in the 2+2 division is n-Inp
‘(were) standing’ modified by the prepositional phrase i Dyin
‘above Him’.

According to Wickes’s principle of continuous dichotomy,
the accents, like Saadya’s translation, also reflect a3 +1 grouping
of this phrase. The entire phrase is governed by the accent zagef
on the word i. The primary division within zagef’s domain is
indicated by the accent pasta on the word >yinn, since there is no
revia‘ present.”°? The result of this division is that * is set apart
from the words Dyian | ory DDT.

252 1 Kgs 7.3a offers an example of >via with pasta preceded by revia‘
(i.e., one of the ways in which a 2+2 grouping could have been ac-
cented in Isa. 6.2). As a result, vin forms a compound modifier with
what follows: ovHAT->Y WR Avoya-dy Opin Ika 1p01 ‘And it was
covered with cedar above the sides which were on pillars.’

253 My analysis here is contra Ben-Shammai’s (2003, 38). Later in the
passage, Saadya mentions that if the sense requires it, 3#75n takrij ‘devi-
ation’ is permissible (Zucker 1984, 19, In. 21), and Ben-Shammai counts
this verse as one of those instances. That is, he observes Saadya group-
ing the phrase 3+1 and says this is an example of takrij since *9 Dvian
are “joined by the accents.” There are two problems with this analysis.
First, for these two words to be joined together by the accents a revia‘
must be present before the pasta. Second, Saadya discusses takrij only
after he cites Isa. 6.2. This suggests that he chooses Isa. 6.2 to illustrate
the salient point of the passage—that the mashir ‘commonly accepted
sense’ of a passage is to take it according to the proper arrangement of
its words; that is, he did not choose Isa. 6.2 to illustrate the exception.

398 Habib

Saadya’s explicit reason for this grouping is theological: the
2+2 division of the verse would be blasphemous because “the
angels are set above their Creator” (Qafih 1976, 200, In. 3). Ad-
ditionally, Saadya prefaced this section by saying that word
groupings are determined by the ‘meaning’ (1xyn>x; Qafih 1976,
199, In. 24). Thus, while Saadya’s translation and interpretation
reflect the division of the accents, they are the product, according
to his own words, of theological and semantic considerations.
These considerations also provide a degree of cohesion for the
ideas of dammat and mashur, as, instead of accents, words in the
Bible are grouped according to notions of theologically or seman-
tically correct reading (dammdt). These are the ways in which the
native speakers of Biblical Hebrew/speakers within Saadya’s
community would have understood the obscure passages
(mashur). In order to resolve this ambiguity we should ask
whether the same theological reservations would necessarily re-
quire that the words be grouped in a particular way. I suggest
that this is not the case. Saadya could easily have chosen another
grouping of words to safeguard against a blasphemous interpre-
tation.

The treatment of this same passage by the Karaite exegete
Yefet ben Eli, which is given below, makes it clear that he has
the same theological reservations as Saadya. But his interpreta-
tion of Isa. 6.2 does not reflect the division of the accents. This
indicates that, even given the same reservations, one’s decision
concerning the semantic grouping of words need not match the

grouping reflected by the accents.

Does Saadya Refer to the Accents? 399

fol. 72r

D7 xd Dp aw? 1p)
NT TIN HY RVI
Sy TaD KN

TPINTI VANTIN 70ND
O7DDN TS ONPIT? [a
maw

Ipa “ndS ATMbRA
Dy ws wp7? Dyan
DWN

nT VAD JAX ANI
oxdn vow) 1p ROD

mx oy Sp Span nse
Dx noo pip TIaDdN RT

Doradx pip qn 4p 2
Sy Tiaan Dw 7IRDp
Sombx

As for the phrase xwn 07 xea->y aw
‘Sitting on the throne high and lifted
up’ (Isa. 6.1), it makes known that he
(Isaiah) saw the glory upon

a chair highly lifted up. Ezekiel already

clarified that the chair resembles

the sapphire jewel, as it says 5p”

OWNA~Dy IWR bptd

Nb NNT Va TAX ANN» ‘Above the fir-
mament which was over their heads,
having an appearance like sapphire,
was the likeness of a throne’ (Ezek.
1.26). Now, the phrase orx2n row
Sanne ‘and his hems filled the temple’
(Isa. 6.1), refers to the fact that he saw
the glory above the top of the roof of the
temple [even though it surrounds the
temple because this is the way of every-
thing which is filled from something—

55 over the temple. So,

it is set above it],
it is as if the hems of the glory are over

the temple

754 9099] IOM Ms. A 143 fol. 78r, In. 3 5a°7>x.

55 Added from IOM Ms. A 143 fol. 78r, Ins 3-5 1x9 Da-n>xa non NARI
my dear whe ra nn ww 59 Sad TD.

400

op Tiad Na PoNIM

Habib

and around it. So, he (Isaiah) saw the

xox DroYy 81 8 glory of the Lord [as something] very
DTT? enormous, whereas Ezekiel
Sy THAIN NT
saw the glory on the cherubim. The
Daw ipionnsR 9
Synn po Tay phrase Spina | oryAy Daw
Np ADK ma tb 45 (Isa. 6.2) intends, ‘angels [which]
g
pip TIA2X ONTP 10 stood before the glory above the tem-
Sonbx ple.’
IOM Ms. A 143 fol.
78r
The phrase 15 >yian (Isa. 6.2) is a refer-
ARINWKE nb Svan ip 13 P pian ( )
ence
22 TID OTPN wD to something whose mention came be-
Som DIPNAN ODN 14 fore in the previous verse, and it is
non three
OND TID] NOD NWR 15 things—throne, and glory and temple.
Dyin ip? my XD Now, it is not possible to say ‘above
7 naa on Tx Tiaa55 the glory’ because they (the angels) are
pip <INI {X> 1X7 16 His servants. Nor is it possible that they
"DIDON are over the chair
fol. 78v
pip >> IN pA AYA
with Him. It remains that they are over
OTR NID ITN 1

poyn

the temple. So, the chair was suspended

256 soy OoTpniddx pax 1) ADT OTN wx] RNL Evr. Arab. 1568 fol. 72r,

Ins 15-16 DTpn xa x.

287 >> 8 pa myn] RNL Evr. Arab. I 568 fol. 72r, In. 18 xaix.

Does Saadya Refer to the Accents? 401

nino 1D Domb pip
537 ONTP D'DTWOINI
07258

*ST12298 DNTP NP! 3 standing before the glory.

over the temple in the air and the Sera-

phim were before the feet of the chair,

Yefet’s interpretation of the phrase in question reflects a
2+2 division. He says that the intended sense behind os iw
i Svian | ovTny is ‘angels standing before the Lord above the temple’
(RNL Evr. Arab. I 568 fol. 72r, Ins 9-10; emphasis mine: JH). His
insertion of the prepositional phrase 7i225x oN7p ‘before the
Glory’ makes it clear that the following phrase 52°7>x pip ‘above
the temple’ reflects his understanding of the Hebrew words %yian
i+ as in the same prepositional phrase, viz., closely bound. He
interprets the referent of i as the temple.

Yefet has the same theological reservations as Saadya in
this verse, but resolves them by exegesis in light of the context
and in light of other Scripture. He explains that i >yian ‘above it’
refers to something in the previous verse—either the throne
(xp), the Lord Himself (37), or the temple (52°77). The first two
options are ruled out on theological grounds. The angels can nei-
ther be over the Lord nor His throne, because they would then be
positioned higher than the Lord Himself. This leaves only the op-
tion that the angels are standing over the temple in front of the
Lord (IOM Ms. A 143 fol. 78r, In. 13-fol. 78v, In. 3). Yefet’s exe-
gesis illustrates his understanding of the phrase in light of its con-

text.

258 sin958 ONTP ANI /70795x 537 ONTP O'NIW5NI] RNL Evr. Arab. 1 568 fol.
7 2v, Ins 1-2 9x71 71958 ONTPI NONTP) DANTP DDIW5NI.

402 Habib

The reason Yefet understands the Lord and His angels to be
above the temple in the first place is because of his exegesis in
light of other Scripture. Earlier in the commentary (not shown),
Yefet connects Isaiah’s vision of the Lord with that of Ezekiel’s
(Ezek. 1.16-21). The difference, Yefet says, is that in Ezekiel’s
vision, the Lord is departing from the temple, whereas in Isaiah’s
vision the Lord is dwelling in the temple.”*’ In both visions God’s
throne is on high. This is indicated in Isaiah’s vision by the phrase
xwn on ‘high and lifted up’ (Isa. 6.1) and in Ezekiel’s by yam
which was over their heads, having an appearance like sapphire,
was the likeness of a throne’ (Ezek. 1.26), from the latter of which
Yefet also deduces that the throne in Isaiah’s vision must be made
of sapphire (RNL Evr. Arab. I 568 fol. 72r Ins 3-5). Yefet then
accommodates the phrase D277"nx oN Yow ‘and His hems
filled the temple’ to this scenario: if one thing fills something else,
by definition it is then above it.

Yefet’s exegesis of this passage illustrates that theological
considerations do not force one to read the words as grouped in
a particular way. Both Saadya and Yefet express the same theo-
logical reservations regarding this verse. Yefet resolves this ten-
sion by considering the context and other Scripture. Saadya re-
solves it by recourse to the dammat of the verse. This gives further

support to the idea that Saadya’s conception of dammat ‘grouping

5° The grounds for Yefet’s claim that the Lord is departing the temple
in Ezekiel’s vision are the presence of the 0°35ix ‘wheels’ (see Ezek. 1.15-
21), which Yefet interprets as depicting God on a 41979 ‘chariot’ (RNL
Evr. Arab. I fol. 71v, Ins 3-8).

Does Saadya Refer to the Accents? 403

of words’ may very well be determined by the grouping of the

accents.

2.2.1. Takrij

The next section of SIP concerns the term takrij ‘deviation’. After
Saadya illustrates how the mashur ‘commonly accepted’ sense of
a passage may be understood according to its dammat ‘grouping
of words’, he says that if there is need for ‘deviation’ from this
principle in order to make the meaning clearer, then this is per-
missible (Ins 21-22). Saadya defines takrij as separating that
which is joined and joining that which is separated (In. 22). He
then offers two examples that respectively illustrate these two
processes—Gen. 3.22 and Exod. 35.34. The crucial task for the
present discussion is discerning the basis for the exegetical tradi-
tion from which Saadya deviates. If Saadya’s exegesis of these
verses reflects a deviation from the clear division of the accents,
then this makes it even more probable that the idea of dammat

indeed refers specifically to the accents.

(2) 7B) NAY yyy aie nw T? awa THN Ty ONT 1D Dy TIT | WBN

20292 Max) ONIN pun Da hpy1 Hr nw

‘Then the LORD God said, “Behold, the man has become like

one of us in knowing good and evil. Now, lest he reach out

his hand ‘and take also of the tree of life and eat, and live
forever.”’ (ESV, Gen. 3.22)

404

Habib

RNL Evr. Arab. II C 1 fol. 6r

Line

ARY TP OTR NTIN Td5x Oxp on

swdx1 25x TaIYA TIN TNNID

10 TOD TT TA Nd TNdKI

xm Soxn Rv OpMdy Tow

ands vy

Then God said, ‘Behold! Adam
has become

as one from whom (is) the
knowledge of good and evil.
And now, lest he stretch forth
his hand and take from

the tree of life as well, and eat,
and live

for eternity.’

Zucker (1984, 79)

JNND VT OTNA IA TANAN AdIpI
mag qwo yn aw nyt una
prin

TRY TD AIS 171 O78 “yp Titbs
1Y NUNN [RII DDI TNNOND
70 THTTNON

pn amv ah tpa myn
TNR pT aw nyt una Inxs
moss yaw 7a

qwbxi 15x Any?

Now the Blessed One’s phrase
aiy nvT? 132 THN M7 BTN 1D
yu (Gen. 3.22) reveals to us
the reason (lit., ‘place’)

for the rebuke of Adam. It is
because [it is as if the Lord
said] ‘He (Adam) himself has
become like a master of him-
self. And he (Adam) was no
longer in need of additions
from among

My teachings.’ So, it required
his banishment. AN 7nXd
yi aiv nyt) (Gen. 3.22) is lit-
erally translated as, ‘Like one
who by himself

knows good and evil.’

Does Saadya Refer to the Accents? 405

The question here is how to understand the words 7nx>
139. The accents join these two words together. This is indicated
by the conjunctive accent munah on the word 7nxX>. The word-
unit 1347 is divided from what comes after by the disjunctive ac-
cent zagef. According to the division of the accents, the following
form, nyt) ‘knowing’ must be analysed as an infinitive which
elaborates on the immediately preceding clause.”®° The accents,
therefore, reflect an interpretation such as the following: ‘Behold!
The man has become like one of us (in what way?), knowing good
and evil.’*°! Any takrij ‘deviation’ from the division of the accents
must divide between the two words 12/7 7NX> in some way.

Indeed, Saadya’s translation and commentary reflect an in-
terpretation based on a division between 7nx> and 137. The He-
brew form 1377 may be analysed as the (doubled) preposition 1
‘from’ plus either the 3Ms suffix (= ‘from him’) or the 1cPL suffix
(= ‘from us’). The interpretation which results from the joining
of the two words by the accents requires this to be understood as
‘from us’ (e.g., ESV and LXX). Saadya, however, translates ac-
cording to the former—n1 minhu ‘from him’. He then translates
the Hebrew form np? as a noun—naryn maTifa ‘knowledge’-—
that is in construct with what follows (1w5xi 25x; In. 14). This
suggests that Saadya analysed the Hebrew yy) xiv nyt? 13797 as an

260 See Joiion and Muraoka (2006, 407) for this function of the preposi-
tion -5.

61 This corresponds to the LXX translation of this verse: xai eimev 6 bed¢
Tdobd Adap yéyovev we cic 2& quay tot ywawoxew xaddv xalt movypdv ‘And God
said, “Behold! Adam has become like one of us, knowing good and
evil”’. The genitive article followed by an infinitive modifies the mean-
ing of a previous verb. See Smyth (1920, 82032).

406 Habib

asyndetic relative clause modifying the head 7nx ‘one’ (in the
phrase 7nx3). In the relative clause qw5x1 72x naan ‘knowledge
of good and evil’ is the compound subject and 73n ‘from him’ is
the predicate (as I have translated above in In. 14); the
knowledge of good and evil comes from man himself, with no
outside intervention. His translation is therefore a deviation
(takrij) from the division of the accents.

Furthermore, Saadya’s commentary suggests that his takrij
in Gen. 3.22 is a deviation from the accents. He says that God
rebukes Adam, because Adam has become like one who already
knows everything (TNnox2 lit. ‘a teacher/master’; In. 20). Adam
no longer needs God to teach him good and evil (Ins 20-21). This
interpretation reflects an understanding of 7nX>d as not bound to
what follows. Saadya offers his own literal translation (pn) in
what follows (Ins 21-22), which confirms the above analysis of
his understanding of the Hebrew syntax—tnn3 is followed by an
asyndetic relative clause. This is indicated by the fact that Saadya
inserts the Arabic relative pronoun 7A man in his literal transla-
tion—Adam is like one who (jf man) knows good and evil by
himself. In both syndetic and asyndetic relative clauses in Biblical
Hebrew, the word being modified is usually set off from its mod-
ifier by a disjunctive accent.*” This strongly suggests that Saadya
interpreted a break between 1nx> and 1347, contra the division of

the accents.

°° E..g., asyndetic: :17Aw» °277 “wai ‘Happy are those who keep My ways’
(Prov. 8.32); syndetic: ...7ay ANNI WWxX AWRA DINA WX" ‘And the man
said, “The woman which you gave to be with me...” (Gen. 3.12).

Does Saadya Refer to the Accents? 407

It seems, therefore, that in Gen. 3.22 the takrij ‘deviation’
to which Saadya refers is a deviation away from the division of
the accents. The process in this takrij was, as he says in SIP, to
separate that which is joined together. The words that are joined
together are 139 Tnx. In this particular phrase both the vocali-
sation and the accents join these words together—rnx is the form
normally used in construct relationships whereas 7nx is typically
the absolute form.”°? Using the example of Gen. 3.22 alone, it is
not certain whether the division from which Saadya deviates is
that based on the accents or the vocalisation. The next example
of takrij that Saadya gives strongly supports the interpretation of

the term as referring to deviation from the accents.

(3) PTs TAS. OTPwWA OPI AVIIS TAN
‘On the Menorah shall be four cups. Its knobs and its flow-
ers shall be almond-shaped’ (Exod. 25.34/37.20)

This verse is one of yan yad px ATINA Mixjpn wn ‘five verses
in the Torah for which no decision has been reached’ (b. Yoma
52a.10-52b.1). The question here is whether the word opwa
should be grouped with what comes before or after. The former
yields the translation ‘Four almond-shaped cups [shall be] on the
Menorah.’ The latter produces the translation I have offered
above (Breuer 1989, 369; Kogut 1994, 35). All the major Tiberian
codices group 0°1pwW/” with what comes after by placing an ’atnah

on the previous word 0°33, thereby grouping D'pwn with what

63 Tbn Ezra draws attention to the significance of the vocalisation of
this phrase (Kogut 1994, 41-42). Al-Fasi’s also considers the difference
between these forms to be one of contextual status (Skoss 1936, I:61-
62).

408 Habib

comes after. Morphosyntactically, both options are possible.
Grouped with what comes before, o°TpW” is an MPL attributive
adjective modifying MPL 0’'v13. Grouped with what comes after,
o’Tpwn may be analysed as the predicate of n751 AAD3, both
MPL. Therefore, if Saadya’s translation reflects a division whereby
oTpwn is grouped with what is before (0’p23), then that which
Saadya is deviating from—the commonly accepted sense

(mashur)—must be the division of the accents.

RNL Evr. Arab. II C 1 fol.
175v (25.34)

VIAN TINIAN nNT DI And as for the Menorah, four al-
MNNNI Le

Lines

mond-shaped cups—
it is also said ‘vessels’—and its
xoamaxani nxndnawria dp 1g
apples
NJDNI01—=—-19 ~~ and its lilies.
RNL Evr. Arab. II C 1 fol.
214v (37.20)

YIN TININON NNT"D} ~=17 And as for the Menorah, four

Lines

almond-shaped cups and its ap-
NOTM|aNANT NNN NRANI ~=—-18

ples

NAIONIOI 19 and lilies.

I have found no extended commentary on these verses, but
Saadya’s translations reflect a division whereby no pwn is
grouped with what comes before it, suggesting that the takrij ‘de-
viation’ contrasts with the division of the accents. This corre-
sponds to the second kind of takrij mentioned in SIP above—join-

ing together that which is separated.

Does Saadya Refer to the Accents? 409

Saadya’s translation differs from the division of the accents
in terms of (1) the grammatical agreement of o pW orpay in the
Arabic translation and (2) the addition of an extra -1 ‘and’ con-
junction to the Arabic translation of 7"7n52. Saadya’s translation
is virtually identical in both instances. The lone exception is the
addition of another gloss for the word n’pal3 ‘cups’-—ox11 ‘vessels’
in his translation of 25.34.7°* Saadya translates the Biblical He-
brew masculine noun phrase op opr; into the Arabic femi-
nine plural phrase nxndbn nxnsi. Before the word xampxan ‘its
apples’ (= a9h53) Saadya adds the conjunction waw, setting it
apart from what precedes. I agree with Kogut (1994, 143) here,
who says that the sense of Saadya’s translation seems to be that
the Menorah contains three things: (1) four almond-shaped cups,
(2) its apples, and (3) its lilies.*°° In any case, this is clearly an
example of takrij as y>xn ta‘lif ‘joining’ (SIP In. 24).

Since Saadya listed Gen. 3.22 and these verses in Exodus as
examples of takrij, both examples must be deviating from the
same thing—the accents. Gen. 3.22 is an example of separating
that which is joined together (SIP, In. 22) and the Exodus pas-
sages are examples of joining together that which is separated
(SIP, Ins 22, 24).

264 See Ratzaby (1998, 340 Ins 14-15) for Saadya’s explanation for this
alternate gloss.

65 Kogut’s description of Saadya’s approach must be slightly refined for
the sake of accuracy. He claims that Saadya here chooses the tradition
of division of this verse which was ‘rejected’ (nnn). It is clear from the
present discussion, though, that Saadya is not ‘choosing’ a tradition, but
rather deviating (takrij) from one.

410 Habib

3.0. Conclusion

The above analysis suggests that the accents are indeed what
Saadya is referring to in SIP as a means of clarifying obscurities.
While Saadya does not explicitly mention them, the accents
strongly correlate with every key part of the description. Saadya’s
use of the term dammat in Gen. 6.18 points to a particular group-
ing of words in that verse according to the Oral Law. Saadya’s
exegetical principle of recourse to the Oral Law where the zahir
‘apparent meaning’ is unsatisfactory raises the question of what
the gahir (i.e. default) grouping of words was for Gen. 6.18. The
simplest answer is the accents, since they divide the verse in the
way which Saadya was trying to avoid. The way in which the
original users of Biblical Hebrew/those in Saadya’s community
would have understood a verse (mashur) facilitates the arrange-
ment of words into groups (dammat). But what feature of the text
purports to reflect the word groupings of biblical verses as they
were originally understood? The accents are the best candidate.
This original manner in which words were arranged must be
maintained so that what is joined together in ‘the reading’ (a/-
qgird’a) will be joined in meaning. We saw that the term a/-qird’a
has a special meaning of reading with chant or oral performance,
since Saadya used the word al-tilawa to refer to reading without
chant or, perhaps, plain reading. How does a reader know what
to chant/perform? The best candidate is the accents. When
Saadya illustrated the correct use of this interpretive device in
Isa. 6.2, the proper arrangement of words reflected the division
of the accents. When giving an example of takrij whereby one

must break apart that which the mashur joins together, Saadya

Does Saadya Refer to the Accents? 411

gave Gen. 3.22. Here Saadya interpreted as separate what the ac-
cents join together. For the example of takrij whereby one must
join what the mashur separates, Saadya gave Exod. 25.34. There,

he joined together what the accents separated.

References

Allony, Nehemya, ed. 1969. Ha-Egron: Kitab Usil al-Sif al-TIbrani
by Rav Se‘adya Ga’on. Jerusalem: The Academy of the He-
brew Language. [Hebrew]

Ben-Shammai, Haggai. 1991. ‘Saadya’s Introduction to Isaiah as
an Introduction to the Books of the Prophets’. Tarbiz 60/3:
371-404. [Hebrew]

. 2003. ‘The Tension between Literal Interpretation and

Exegetical Freedom: Comparative Observations on Saadia’s
Method’. In With Reverence for the Word: Medieval Scriptural
Exegesis in Judaism, Christianity, and Islam, edited by Jane
Dammen McAuliffe, Barry D. Walfish, and Joseph W. Goe-
ring, 33-50. Oxford: University Press.

Breuer, Mordechai. 1989. The Accents of the Bible in the Twenty-
One Books and in the books of Job, Proverbs, Psalms. Jerusa-
lem: Horev. [Hebrew]

Brody, Robert. 1998. The Geonim of Babylonia and the Shaping of
Medieval Jewish Culture. New Haven, CT: Yale University

Press.

. 2006. Rav Saadya Gaon. Jerusalem: The Zalman Shazar
Center. [Hebrew]
. 2013. Sa‘adyah Gaon. Translated by Betsy Rosenberg.

Portland, OR: The Littman Library of Jewish Civilization.

412 Habib

Diaz Esteban, Fernando. 1975. Sefer ?Oklah Wé-’Oklah: Coleccién
de Listas de Palabras Destinadas a Conservar La Integridad Del
Texto Hebreo de La Biblia Entre Los Judios de La Dad Media.
Textos y Estudios ‘Cardenal Cisneros’ 4. Madrid: Consejo
Superior de Investigaciones Cientificas.

Dotan, Aaron. 1967. The Diqduge ha-Te‘amim of Aharon ben Moshe
ben Asher: With a Critical Edition of the Original Text from
New Manuscripts. 3 vols. The Academy of the Hebrew Lan-
guage Texts and Studies 7. Jerusalem: The Academy of the

Hebrew Language. [Hebrew]

. 1970. ‘Prolegomenon’. In Two Treatises on the Accentua-
tion of the Old Testament: Ta‘ame Emet on Psalms, Proverbs,
and Job; Ta‘ame Kaf-Alef Sefarim on the Twenty-Only Prose
Books, by William Wickes. New York: KTAV Publishing
House, Inc.

. 1996. ‘Masoretic Schools in the Light of Saadya’s Teach-
ing.’ In Proceedings of the Twelfth International Congress of

the International Organization for Masoretic Studies, edited by
E. J. Revell, 1-9. Scholars Press.

. 1997. The Dawn of Hebrew Linguistics: The Book of Ele-
gance of the Language of the Hebrews by Saadia Gaon. 2 vols.

Jerusalem: World Union of Jewish Studies. [Hebrew]
. 2005. The Awakening of Word Lore: From the Masora to the
Beginnings of Hebrew Lexicography. Jerusalem: The Acad-

emy of the Hebrew Language. [Hebrew]
Gil, Moshe. 1992. A History of Palestine, 634-1099. Translated by
Ethel Broido. Cambridge: Cambridge University Press.

Does Saadya Refer to the Accents? 413

de Goeje, Michal Jan (ed.). 1894. Kitab al-Tanbih wal-’Ishraf by
Abii al-Hasan °Ali ben al-Husdayn ben ‘Ali al-Mas di. Biblio-
theca Geographorum Arabicorum 8. Leiden: Brill.

Habib, Joseph. 2021. ‘Accents, Pausal Forms and Qere/Ketiv in
the Bible Translations and Commentaries of Saadya Gaon
and the Karaites of Jerusalem’. PhD dissertation, University
of Cambridge.

Hitin-Mashiah, Rachel. 2011. ‘Biblical Accents in the Preface of
the *Egron (Thesaurus) by Rav Se‘adya Ga’on’. In Israel: Lin-
guistic Studies in the Memory of Israel Yeivin, edited by Rafael
I. Zer and Yosef Ofer, 113-128. Jerusalem: The Hebrew
University Bible Project. [Hebrew]

de Hoop, Raymond. 2008. ‘Stress and Syntax, Music and Mean-
ing: The Purpose and Function of the Masoretic Accentua-
tion System’. Journal of Northwest Semitic Languages 34/2:
99-121.

Jotion, Paul, and Takamitsu Muraoka. 2006. A Grammar of Bibli-
cal Hebrew. Subsidia Biblica 27. Rome: Pontifical Biblical
Institute Press.

Kogut, Simcha. 1994. Correlations between Biblical Accentuation
and Traditional Jewish Exegesis. Jerusalem: Magnas Press.
[Hebrew]

Malter, Henry. 1921. Saadia Gaon: His Life and Works. Philadel-
phia: Jewish Publication Society of America.

Ofer, Yosef. 2019. The Masora on Scripture and Its Methods. Fontes
et Subsidia ad Bibliam pertinentes 7. Berlin: De Gruyter.

Ognibeni, Bruno. 1995. La Seconda Parte del Sefer ’Oklah
We?’Oklah: Edizione del Ms. Halle, Universitdtsbibliothek Y B

414 Habib

40 10, Ff. 68-124. Textos y Estudios ‘Cardenal Cisneros’ 57.
Madrid: Instituto de Filologia del CSIC, Departamento de
Filologia Biblica y de Oriente Antiguo.

Park, Sung Jin. 2014. “Pointing to the Accents in the Scroll”:
Functional Development of the Masoretic Accents in the
Hebrew Bible’. Hebrew Studies 55: 73-88.

Penkower, Jordan S. 1981. ‘Maimonides and the Aleppo Codex’.
Textus 9/1: 39-128.

Pitcher, Sophia Lynn. 2020. ‘A Prosodic Model for Tiberian He-
brew: A Complexity Approach to the Features, Structures,
and Functions of the Masoretic Cantillation Accents.’ PhD
dissertation, University of the Free State.

Polliack, Meira. 1997. The Karaite Tradition of Arabic Bible Trans-
lation: A Linguistic and Exegetical Study of Karaite Transla-
tions of the Pentateuch from the Tenth and Eleventh Centuries
C.E. Leiden: Brill.

Qafih, Yosef (ed. and trans.) 1966. Tehilim ‘im Tirgum u-Ferush ha-
Ga’on Rabbenu Se‘adya ben Yosef Fayumi. Jerusalem: Amer-
ican Academy of Jewish Research.

(ed.). 1969. Kitab al-Mukhtar fi al-Amdndt wal-I‘tiqadat

(Sefer ha-Nivhar ba-’Emunot u-va-De‘ot). Jerusalem: Sura.

(ed. and trans.). 1976. Mishle im Tirgum u-Ferush ha-Ga’on
Rabbenu Se‘adya ben Yosef Fayumi. Jerusalem: Ha-Va‘ad le-
Hotsa’at Sifre Rasag.

Ratzaby, Yehuda. 1993. Tafsir Yesha‘ya le-Rav Se‘adya. Kiryat
Ono: Mekhon Moshe.

. 1998. Sefer Perushe Rav Se‘adya Ga’on le-Sefer Shemot:

Magor ve-Tirgum. Jerusalem: Mosad ha-Rav Kook.

Does Saadya Refer to the Accents? 415

Revell, E. J. 1974. Review of Seadya Gaon, “Ha-Egron: Kitab Usul
as-Shi’r al-’Ibrani”. Journal of Semitic Studies 19: 123-29.

Schechter, Solomon. 1901. ‘Geniza Specimens.’ The Jewish Quar-
terly Review 14/1: 37-63.

Shoshany, Ronit. 2009. ‘The Biblical Accents’ Original Purpose’.
In Mas?’at Aharon: Linguistic Studies Presented to Aaron Dotan,
edited by Moshe Bar-Asher and Chaim E. Cohen, 469-86.
Jerusalem: The Bialik Institute. [Hebrew]

Skoss, Solomon L. 1936. The Hebrew-Arabic Dictionary of the Bible
Known as Kitab Jami‘ al-Alfaz (Agron) of David Ben Abraham
al-Fasi the Karaite (Tenth Cent.). 2 vols. Yale Oriental Series
Researches 20. New Haven, CT: Yale University Press.

Smyth, Herbert Weir. 1920. A Greek Grammar for Colleges. New
York: American Book Company.

Sokolow, Moshe. 1984. ‘Saadiah Gaon’s Prolegomenon to
Psalms’. Proceedings of the American Academy for Jewish Re-
search 51: 131-74.

Stern, Sacha. 2019. The Jewish Calendar Controversy of 921/2 CE.
Leiden: Brill

Vollandt, Ronny. 2015. Arabic Versions of the Pentateuch: A Com-
parative Study of Jewish, Christian, and Muslim Sources. Bib-
lia Arabica 1. Leiden: Brill.

Wickes, William. 1887. A Treatise on the Accentuation of the
Twenty-One So-Called Prose Books of the Old Testament. Ox-
ford: Clarendon Press.

Yeivin, Israel. 1959. ‘The Accentuation of Rabbinic Texts with
Te‘amim’. LéSonénu 24/1-2: 47-69.

416 Habib

Zewi, Tamar. 2015. The Samaritan Version of Saadya Gaon’s Trans-
lation of the Pentateuch: Critical Edition and Study of MS Lon-
don BL OR7562 and Related MSS. Biblica Arabica 3. Leiden:
Brill.

Zucker, Moshe, ed. 1984. Perushe Rav Se‘adya Ga’on li-Vreshit.
New York: Bet ha-Midrash le-Rabanim ba-Ameriqa.

INDEX

accents, x, xxi, 11-12, 14, 52,

66, 182, 243, 245, 263, 279,

289, 295, 297, 306, 318,
330, 336, 340, 342-45,
361-62, 377-78, 381, 384,
386-91, 393-94, 397-98,
403, 405-10

accentuation, xix, xxi, 1-2, 5,
7, 9, 11-12, 14, 126, 174,

176, 182-83, 197, 289, 307,
326, 329-30, 337, 344, 349,

361

-alef, 163, 167, 169-70, 183-
84, 197

Alsheikh, Moses, xxi, 330,
345-54, 356-64

Arabic, v, ix—x, xviii, 24, 117-

20, 125, 146, 156, 158, 189,
203-4, 214-17, 237, 239,
240-42, 290-91, 293, 378,
382, 391, 406, 409, 414-15

?atnah, 267, 275, 280-81, 303-

17, 322-23

-atnahta, 294
‘ayin, 164, 170-71, 173
-azla, 294

Babylon, 23, 145, 254, 335-
36, 379

Babylonian Masora, xvi, 2, 19,
24-25, 40, 46, 65, 72-73

Babylonian te‘amim, 289

begadkefat, 243

Byzantine, v, xvii, 112, 163,
165-67, 170-71, 173, 181,
183-85, 187, 190-96, 198,
337

cantillation, xi—xii, xix, xxi, 31,
243-48, 251-55, 257-62,
264-66, 269-73, 275-81,
283-85, 289, 295-96, 329-
30, 334, 340, 342-43, 345,
352, 356, 359, 364, 393

codex, xi—xii, xv—xvi, 1, 5, 7,
9-10, 12-14, 24, 26-29, 53,
56, 66, 71, 75-80, 83, 86,
90, 92, 95, 96, 98-99, 103-
4, 106-9, 163-67, 183, 185,
189, 191, 194, 197, 205,
213, 236, 238-39, 291,
294-96, 334, 407

codicology, viii, xviii, 84, 164-
65, 184, 187, 189-90, 192,
197-99

418

conjunctive accent, 58, 176-
77, 181-82, 243, 245, 294-
95, 315, 317, 343, 405

construct, 245-50, 252-58,
260, 274-75, 279, 282-83,
405, 407

dagesh, xvii, 126, 163-64, 169,
170, 172-81, 183-84, 197,
214, 242

dammat, 384, 386-90, 393,
398, 402-3, 410

darga, 294, 357

defective, 35, 39, 4-44, 62, 67,
69, 99, 125, 147

dehi, 314-17

derash, xxi, 330, 337-38, 343

disjunctive accent, 176-78,
182, 243, 245, 264, 275,
278-81, 294-95, 303-7,
309, 312-15, 319, 343, 353,
356, 360, 387, 405-6

Extended Tiberian, xvii, 197-
98, 201

Firkovitch collections, 79, 204

galgal, 294

geresh, 257, 292, 294

guttural, 55, 125, 171, 173,
179, 183, 196-97, 214, 221,
224-25, 227, 232, 234, 236

hatef, xviii, 30, 173, 188-90,
205, 207-9, 211, 219, 234-
36, 238-39

het, 54, 164, 170-73

hireq, 54-55, 62, 167, 185,
186, 188-89, 235

holem, 52

iconicity, 128

imperfect performance, xviii,
204

interchange, 135, 137, 146,
165, 172, 184-90, 199, 206,
215, 217, 237

interpretation, x—xi, xv, xix, 2,
4-5, 47-48, 116, 148, 155,
174, 244, 247, 250, 254,
256-62, 266-68, 281, 283,
285, 329, 331-34, 336-37,
340-42, 344-46, 348-49,
353-54, 356, 358, 360-63,
383, 385-88, 390-91, 393,
396, 398, 401, 405-7

Karaite, v, x, xviii, 18, 203-5,
213, 236, 239, 240-42, 291,
325, 398, 413-15

ketiv, v, xvi, xvii, xix, 75-76,
78, 80, 83, 98, 102-4, 106-
7, 110, 115-33, 135-36,
138-39, 141-42, 144-54,
204, 207, 210, 234, 243-44,
284

Studies in the Masoretic Tradition of the Hebrew Bible 419

lamed, 51, 172, 175, 178-80,
182-83

legarmeh, 294

mappiq, 170, 277-78, 380

magqgef, 14, 306, 316, 335

mashur, 389-90, 393, 397-98,
403, 408, 410

masora magna, 3, 24, 41, 76,
81, 192, 334, 336

masora parva, xvi, 3, 40-41,
76, 78, 80-81, 84, 95-101,
106-7, 109-10, 145, 192

Masoretes, vii, xi, xix, 14, 18,
204, 236, 244, 273, 275,
278, 281, 284, 299, 311,
318, 329, 372, 378, 380-81

Masoretic Text, 1, 19-20, 118,
218, 243, 284, 300, 329,
333

mayela, 294, 303

merkha, 14, 294, 317

midrash, vi, X, Xx, 329-32, 334,
336, 338, 340, 345, 349,
358, 362, 364-65, 367, 371,
374-76, 416

mille‘el, 294

millera‘, 294

Nehardea, 28, 73

nequddot, 292-93, 320

niqqud, 243, 292

nun, v, xvi, 75-77, 80, 84, 86,
92-96, 100, 105, 110, 134,
175-76, 178-80, 182-83

palaeography, vii, viii, 82,
185, 187, 191-92, 194,
196-99

Palestinian te‘amim, 300, 320

Palestinian Yeshiva, xi, 204

Palestino-Tiberian, xvii, 164—
65, 197

parallelism, 254, 295, 305,
307-8, 310, 320, 321

parasha, 57-59, 61, 333, 351

parashiyyot, 57

paratextual, 84-85, 92, 108-9,
362

pashta, 8, 306

patah, 30-31, 52, 172-73,
185-86, 188-90, 206-12,
214-19, 221, 225, 227,
229-30, 232, 234, 237, 260

pausal, x, 15, 243, 245, 263-
65, 267-69, 271, 285-86,
304, 330, 413

pazer, 294, 352-53, 361

Pentateuch, vi, xvii, xxi, 19,
24-25, 71-72, 83-84, 95,
100-1, 112, 118, 144, 156-
57, 377, 382, 414-16

peshat, 330

pesiq, 338

420

pisqa, 47, 57-60

Piyyut, 292

plene, 6, 33-40, 42-45, 47-49,
55, 63, 65, 67- 68, 146, 151

poetry, xx, 295-96, 303, 307,
309, 313-14, 321, 326, 378

puncta extraordinaria, 331

games, 185-86, 188-89, 206,
208, 211-12, 214-18, 224-
25, 229, 233-34

gere, V, XVi-xvii, xix, 11, 75-
76, 78, 80, 83, 96-99, 101-
4, 106, 107, 110, 115-20,
122-32, 135-36, 139, 141-
42, 144-45, 147-54, 204,
244, 284, 331

gir@a, 390-93, 410

rafe, 30, 126, 163, 169-70,
199

resh, 6, 175, 179-80

revia‘, 294, 314-16, 356, 360-
61, 397

revia‘ mugrash, 314-16

Ruth, 19, 298, 346, 348-55,
357-61, 363, 368, 372, 376

Saadya Gaon, vi, x, xxi, 377-
82, 384-91, 393-94, 396-
98, 401-3, 405-13, 416

Samaritan, v, viii, xvi, 115-18,
121, 123, 125, 127, 133-34,
136-38, 140-41, 143-44,

Index

147, 149, 151-54, 156-58,
160, 366, 373, 416

samekh, 63, 175-76, 178

segol, 6-7, 55, 185-86, 188-
89, 206-9, 211, 213-14,
216-18, 223-24, 228, 232,
254, 271, 335

sere, 185-86, 188-89, 213

sevirin, 51

shalshelet, xxi, 329-30, 337-
43, 345-48, 361, 363, 368

shewa, xvii-xviii, 163-64, 170-
73, 175, 180-81, 183-86,
188-90, 197, 205, 219, 227,
229, 234-36, 238-39, 260-
61, 271, 380

shureq, 189

silluq, 275, 294, 303-5, 307,
312, 314-16, 386

siman, 6-7, 60, 242

sof-pasuq, 30

stress, xv, 7-11, 15, 275, 289,
294, 306, 315-17

suppletion, 128, 134

Sura, 28, 73, 276, 378, 414

syllable, xviii, 10-11, 14, 119,
121, 124, 166, 169, 171-74,
176-77, 179, 181, 183-84,
186, 196-97, 205, 213, 225,
234-36, 238-39, 243, 304-
5, 315-17, 330, 343, 356

Studies in the Masoretic Tradition of the Hebrew Bible 421

takrij, 397, 403, 405-10

tarha, 14

taw, 8, 179

te‘amim, 243, 289-90, 292-93,
294-97, 299, 301-9, 311-
15, 317-21, 329-30, 336,
338-39, 342-43, 345, 349,
353, 357, 361-63

telisha, 9-10, 294, 335-36

teres, 294

tevir, 257, 279-80, 294, 322

Tiberian pronunciation
tradition, 55, 179, 204,
213-14, 234

Tiberias, xi-xii, 23, 108, 204,
379

tifha, 294, 303, 305, 307, 312-
14, 322

tosafot, 297

transcriptions, xviii, 203-5,
239

verse length, 297-99, 301-2

vocalisation, xiii, xv, xvii—xix,
1-2, 5, 7, 9, 31, 49, 53-55,
62-63, 80, 104, 108, 115,
126, 150, 164, 169, 171,
185, 190, 192, 196-99,
205-6, 213-14, 218, 236,
239, 243-48, 251-54, 257,
259-63, 265, 269, 271, 274,
276-78, 281, 283-85, 407

waw consecutive, 9

yetiv, 294, 306

yod, 234-35

zagef gadol, 353, 359-61

zagef qatan, 264, 267, 279-80,
294, 304-7, 309, 311, 313,
318, 322-23, 349, 353,
359-61, 386, 397, 405

zarqa, 247, 294

zayin, Vv, xvi, 75-76, 78, 80,
84, 86, 92-96, 100, 105,
110, 175-76, 184

- Cambridge Semitic :

iia Saati? | SEPM
& aed > ‘> b Ee oe "| on ‘ rcs

About the series

This series is published by Open Book Publishers in collaboration
with the Faculty of Asian and Middle Eastern Studies of the University
of Cambridge. The aim of the series is to publish in open-access
form monographs in the field of Semitic languages and the cultures
associated with speakers of Semitic languages. It is hoped that this
will help disseminate research in this field to academic researchers
around the world and also open up this research to the communities
whose languages and cultures the volumes concern. This series includes
philological and linguistic studies of Semitic languages, editions of
Semitic texts, and studies of Semitic cultures. Titles in the series will
cover all periods, traditions and methodological approaches to the field.
The editorial board comprises Geoffrey Khan, Aaron Hornkohl, and
Esther-Miriam Wagner.

This is the first Open Access book series in the field; it combines the high
peer-review and editorial standards with the fair Open Access model
offered by OBP. Open Access (that is, making texts free to read and
reuse) helps spread research results and other educational materials to
everyone everywhere, not just to those who can afford it or have access
to well-endowed university libraries.

Copyrights stay where they belong, with the authors. Authors are
encouraged to secure funding to offset the publication costs and thereby
sustain the publishing model, but if no institutional funding is available,
authors are not charged for publication. Any grant secured covers the
actual costs of publishing and is not taken as profit. In short: we support
publishing that respects the authors and serves the public interest.

UNIVERSITY OF
CAMBRIDGE
Faculty of Asian and Middle

Eastern Studies

You can find more information about this serie at:
http://www.openbookpublishers.com/section/107/1

Other titles in the series

Diachronic Variation in the Omani Arabic Vernacular of
the Al-‘Awabi District
From Carl Reinhardt (1894) to the Present Day

Roberta Morano
https://doi.org/10.11647/OBP.0298

Sefer ha-Pardes by Jedaiah ha-Penini
A Critical Edition with English Translation

David Torollo
https://doi.org/10.11647/OBP.0299

Neo-Aramaic and Kurdish Folklore from Northern Iraq
A Comparative Anthology with a Sample of Glossed Texts, Volume 1

Geoffrey Khan, Masoud Mohammadirad, Dorota Molin & Paul M. Noorlander
https://doi.org/10.11647/OBP.0306

Neo-Aramaic and Kurdish Folklore from Northern Iraq
A Comparative Anthology with a Sample of Glossed Texts, Volume 2

Geoffrey Khan, Masoud Mohammadirad, Dorota Molin & Paul M. Noorlander
https://doi.org/10.11647/OBP.0307

The Neo-Aramaic Oral Heritage of the Jews of Zakho
Oz Aloni

https://doi.org/10.11647/OBP.0272

Points of Contact
The Shared Intellectual History of Vocalisation in Syriac, Arabic,
and Hebrew

Nick Posegay

& Winner of the British and Irish Association of Jewish Studies (BIAJS) Annual
Book Prize

https://https://doi.org/10.11647/OBP.0271

A Handbook and Reader of Ottoman Arabic
Esther-Miriam Wagner (ed.)

https://doi.org/10.11647/OBP.0208

Diversity and Rabbinization
Jewish Texts and Societies between 400 and 1000 CE

Gavin McDowell, Ron Naiweld, Daniel Stékl Ben Ezra (eds)
https://doi.org/10.11647/OBP.0219

New Perspectives in Biblical and Rabbinic Hebrew
Aaron D. Hornkohl and Geoffrey Khan (eds)

https://doi.org/10.11647/OBP.0250

The Marvels Found in the Great Cities and in the Seas
and on the Islands
A Representative of ‘Aga’ib Literature in Syriac

Sergey Minov
https://doi.org/10.11647/OBP.0237

Studies in the Grammar and Lexicon of Neo-Aramaic
Geoffrey Khan and Paul M. Noorlander (eds)

https://doi.org/10.11647/OBP.0209

Jewish-Muslim Intellectual History Entangled
Textual Materials from the Firkovitch Collection, Saint Petersburg

Camilla Adang, Bruno Chiesa, Omar Hamdan, Wilferd Madelung, Sabine
Schmidtke and Jan Thiele (eds)

https://doi.org/10.11647/OBP.0214

Studies in Semitic Vocalisation and Reading Traditions
Aaron Hornkohl and Geoffrey Khan (eds)

https://doi.org/10.11647/OBP.0207

Studies in Rabbinic Hebrew
Shai Heijmans (ed.)

https://doi.org/10.11647/OBP.0164

The Tiberian Pronunciation Tradition of Biblical

Hebrew

Volume 1

Geoffrey Khan a a
ta?

Winner of the 2021 Frank Moore Cross Book Award for best book related
to the history and/or religion of the ancient Near East and Eastern
Mediterranean

https://doi.org/10.11647/OBP.0163

The Tiberian Pronunciation Tradition of Biblical
Hebrew
Volume 2

Geoffrey Khan

S Winner of the 2021 Frank Moore Cross Book Award for best book related
to the history and/or religion of the ancient Near East and Eastern
Mediterranean

https://doi.org/10.11647/OBP.0194

Studies in the Masoretic
Tradition of the Hebrew Bible

Edited By Daniel J. Crowther, Aaron D. Hornkohl
and Geoffrey Khan

This volume brings together papers on topics relating to the transmission
of the Hebrew Bible from Late Antiquity to the Early Modern period. We
refer to this broadly in the title of the volume as the ‘Masoretic Tradition’.
The papers are innovative studies of a range of aspects of this Masoretic
tradition at various periods, many of them presenting hitherto unstudied
primary sources. They focus on traditions of vocalisation signs and accent
signs, traditions of oral reading, traditions of Masoretic notes, as well as
Rabbinic and exegetical texts. The contributors include established scholars
of the field and early-career researchers.

As with all Open Book publications, this entire book is available to read for
free on the publisher’s website. Printed and digital editions, together with
supplementary digital material, can also be found at www.openbookpublishers.
com.

Cover image: A fragment of a Hebrew Bible manuscript (1 Sam. 25.44—26.8) from the Cairo Genizah
containing vocalisation, accents, Masoretic notes and Masoretic marks (Cambridge University
Library, T-S A8.10). Courtesy of the Syndics of Cambridge University Library.

Cover design: Jeevanjot Kaur Nagpal.

OpenBook iw =)

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Studies in the Masoretic Tradition of the Hebrew Bible"

See other formats